THE CREATIVE ECONOMY THROUGH THE LENS OF URBAN RESILIENCE. AN ANALYSIS OF ROMANIAN CITIES*1

Abstract The creative economy has attracted increasing attention from academia and policymakers for more than two decades. However, despite the fl ourishing literature on this topic, its complex connection with development and its role in strengthening resilience are yet to be properly examined. The paper addresses this issue by investigating how different cities in Romania, with a different intensity of creative industries, have managed to resist and to recover from the aftermath of the Great Recession. Our fi ndings reveal that, as a whole, creative industries strengthen urban resistance against a recession, but do not necessarily fasten urban recovery. As our results suggest, this might be due to the asymmetrical impact across different groups of creative industries. Besides a creative economy proliferation, other factors are also identifi ed as signifi cant resilience drivers. Whilst a better access to healthcare services, higher local investments and a higher decentralization of local budgets appear to enhance the cities’ resistance, higher shares of agriculture and fi nance, as well as a higher income per capita appear to correlate with a faster urban recovery.


Introduction
Since its fi rst development, two decades ago (Howkins, 2001), the creative economy has generated interest for an increasing number of specialists from various fi elds (economics, business and management, law, policy studies, organization studies, geography, sociology and psychology etc.), as well as from world-known organizations, concerned with its potential in promoting growth, prosperity and well-being in regional and national economies (United Nations, 2018a, 2010, 2008; UNESCO, 2013; Dovey and Pra , 2016). At its core, the creative economy comprises economic activities which capitalize creativity through intellectual property rights that form the creative industries. Although there is no generally accepted defi nition regarding the specifi c activities included in this sector, one of the most referred to approaches is provided by the UK Department for Culture, Media and Sport (DCMS), which states that creative industries are 'those industries which have their origin in individual creativity, skill and talent and which have a potential for wealth and job creation through the generation and exploitation of intellectual property' (DCMS, 2001, p. 5).  e importance a ached to creativity within the creative sector is advocated by scholars such as Richard Florida ('human creativity is the ultimate economic resource', Florida, 2002, p. xiii), while others have stressed that 'the industries of the twenty-fi rst century will depend increasingly on the generation of knowledge through creativity and innovation' (Landry and Bianchini, 1995, p. 4). From an economic perspective, the discourse on the creative economy has been continuously diversifi ed, aiming to capture the complex reality in which creativity intertwined with diff erent aspects of growth and development. Consequently, related concepts have also emerged, which linked together the creative economy to cluster theory (Boix et (Doyle, 2015; Tafel-Viia and Lassur, 2013), amongst others.  e proliferation of theoretical and empirical studies on this topic was accompanied by the rising awareness of national and supranational groups, which elaborated policies and strategies with the purpose of supporting the creative sector. In this regard, at the European Union level, for example, several fi nancing instruments were applied for sustaining the cultural and creative sector (Mazilu, 2018, pp. 295-297), among which the Creative Europe programme has received a crucial role (receiving a funding of 1.46 billion EUR for the 2014-2020 period, respectively 1.85 billion EUR for the 2021-2027 period -European Commission, 2018).  ese measures came as a recognition of the importance of the creative economy's place within the European economy, which accounts for 3.8% of the total employment (Eurostat, 2018), while the cultural enterprises, representing 5% of the total number of fi rms in the non-fi nancial business economy, generated 192 billion EUR of added value (Eurostat, 2016).
 e multi-faced spectrum of researches about the infl uence of creative industries within economies ranges from the ones concerning the analysis of the relation be-tween the creative economy and development, including its economic impact (Boix-Domènech and Rausell-Köster, 2018; Hong and Chen, 2017; Boccella and Salerno, 2016), to the ones which focus on revealing the prerequisites that may explain their spatial distribution, formation or dynamics (Kourtit andNijkamp, 2018, 2016;Martinaitytė and Kregždaitė, 2015), as well as the involved workforce (sometimes referred to as the creative class -Florida, 2014;O'Brien et al., 2016). However, while acknowledging that there is a 'virtuous circle' -bidirectional causality -between growth and creative economy (Marco-Serrano, Rausell-Kosterb and Abeledo-Sanchisc, 2014), and that 'regions with high concentrations of creative and cultural industries have Europe's highest prosperity levels' (Power, 2011, p. 5), there is still a lack of evidence dealing with creative industries' capacity to aff ect the anticipating, resisting, and recovering capabilities of the economies (at local, regional and national level) in dealing with various shocks or crises, through resilience.
In the case of Romania, most of the researches analyzing the domestic specifi cities of the creative economy aimed to identify their dimension and spatial distribution (Sava and Leovaridis and Bahnă, 2017). In addition, several national development agencies have conducted studies at regional level (e.g. CIVITTA România, 2019; ADR -Centru, 2016), which along with other case studies (Ișănescu-Ivan, 2018; Asociația Cluj-Napoca 2021 Capitală Culturală Europeană, 2014), off ered a be er picture on the sector at regional or city level. But the Romanian creative industries' relation with resilience remained insuffi ciently addressed, which may be detrimental from a strategic point of view if we take into account that the creative economy has shown a spectacular rise in interest, as well as a sustained growth of turnover in recent years (Network of European Museum Organisations, 2019).
Considering the reasons stated before, our paper aims to examine the capacity of creative industries to act as a bulwark against economic crises. In particular, the paper looks at the Romanian cities over the Great Recession and explores the role that creative industries play in buff ering the economic shock (resistance), as well as their importance for recovery in the a ermath of shocks.
 e remainder of the paper includes the following sections. Section 2 discusses the relevance of considering the urban resilience for deepening the understanding of the creative economy's role in development and it formulates the study hypotheses. Sections 3 and 4 present our methodological approach, the econometric models, and data used. Section 5 discusses the empirical fi ndings, while the last section concludes.

Setting up the lens
As history reveals, the economic growth and development is not a linear, constant process, but its dynamics describe complex pa erns, under the infl uence of the transformations which aff ect human society and which sometimes depict the eff orts to cope with threats, shocks or even crisis, as an expression of what is generally referred to as resilience capacity.  e idea is acknowledged by the economic geography's disciples, as well: ' e fi rst defi ning feature of a 'resilience perspective' on the economic landscape is a recognition that uneven geographical development itself is not some smooth or slowly changing phenomenon, but an inherently shock-prone process, subject to all sorts of disruption, perturbation, and interruption.' (Martin, 2018, p. 4) Although not a new concept, being introduced in the early 1970s, in the fi eld of ecology (Holling, 1973), resilience became a hot topic on interdisciplinary grounds, because of worldwide concerns regarding overcoming adversities, such as natural calamities, economic crises, terrorism, diseases, and other threats which can disturb development and prosperity.
From the diverse array of defi nitions trying to explain and adapt resilience's meaning to particular science fi elds, we can identify approaches that refer it as the amount of disruption that can be absorbed by the system (ecological resilience - Holling, 1973), focusing on keeping functional capacity, while others emphasize on maintaining effi ciency (engineered resilience - Pimm, 1984).  ese show a double hypostasis of resilience -resistance and recovery.
 e acceptations on resilience depend on the level of appliance, as well. Starting from individual level, resilience can be seen as a feature, outcome or process of peoples overcoming crisis situations (Southwick et al., 2014); also, it can be considered at community level, including local, city, regional or national level. As the area of assessment expands, so does the complexity of forces explaining the resilient capacity of a system because this is not always assured by the simple aggregation of the resilience of its smaller components.
From an economic perspective, a special a ention is payed to urban resilience, due to the importance of cities as key engines for growth. It is estimated that by 2030 more than 60 percent of the world population will live in metropolitan areas (United Nations, 2018b), while major cities contribute with signifi cant amounts to their countries' GDPs (e.g. in 2019, Tokyo was the fi rst city in the world, with an estimate of $1.6 trillion GDP -World Economic Forum, 2020). As 'urban resilience can be conceived as a multidisciplinary framework to analyze the reactive, adaptive and transformative capacities of (and within) urban systems' (Olazabal et al., 2012), revealing the factors that explain the shaping of economic resilience at city level could provide clues which will promote sustainability of the cities over time, as they will deal with the ever-changing environment remodeled by a globalized and digitalized world.
Given the increased interest in explaining the resilience capacity of economies, there are only few studies focusing on the role played by creative industries (Buheji, 2019;Pra , 2015). Given that they are accounting for around 7 percent of fi rms, 4 per-cent of employees, and 3 percent of turnover at national level (2008-2017), studying their importance for explaining economic resilience might be worthwhile. A focus on their evolution during the crisis shows no major changes at national level. However, zooming the analysis to the urban level reveals signifi cant changes (please see Figure  1). If before the crisis (2008) Odorheiu Secuiesc, București, Iași, Alexandria, Cluj-Napoca and Timișoara were hosting the highest concentrations of creative industries (measured by the share of enterprises' turnover), it was only Iași and Cluj-Napoca that reported a growth during the resistance period (2008-2011) 1 .  e recovery phase (2011-2017) brought out some essential changes which placed Medgidia, Fălticeni and Turnu Măgurele in the fi rst positions in terms of creatives industries' concentration.
Considering the dynamics of creative industries across the Romanian cities over the last recession, the paper aims to investigate their role in confi guring the economic resilience capacity at city level. Considering previous studies that off er clues on the potential positive impact of the creative economy on resilience (Buheji, 2019;Pra , 2015) and the ability of the sector 'to sail against the tide' (Fontainha and Lazzaro, 2019), our research explores the relation between creative industries and the resistance and recovery of the cities, at the level of Romania.
Several related questions lie at the heart of our research: Given the very diff erent reaction of Romanian cities to the Great Recession, what explains this variation? Are creative industries a signifi cant factor in explaining be er resistance of cities? Are creative industries an important trigger of cities recovery in the a ermath of shocks? Are there any diff erences across the various creative groups in their role for urban resistance and recovery? Which are the other factors that might help cities to be er cope with economic shocks and boost their recovery?
In order to address them, we formulated the following hypotheses to be tested: H1: Creative industries can act as a bulwark against economic shock, being positively associated with a be er resistance of cities; H2: Creative industries can boost recovery in the wake of economic shocks; H3:  e role of creative industries in supporting resistance and recovery diff ers across creative classes (please see Table A1 in Appendix A for more details about the creative industries' classifi cation).
Considering its latest positive results in terms of revenues and employment, but also acknowledging that the Romanian creative economy is in a rather emerging stage, we hope that our research, given its dynamic approach, will provide some pertinent arguments to the policy makers, as well as to the concerned agencies and professionals, to address creative industries with an increased a ention, as the sector may represent a key element in enhancing resilience at the city level. Notes: The dynamics of creative industries' concentration is displayed by using location quotients (in terms of enterprises' turnover). The maps were made with Philcarto, http://philcarto.free.fr

Empirical strategy
For analyzing its connection with the creative economy, our study relies on a widely used measure of resilience which assumes that the trajectory of the national economy as a whole is taken as the expected change of lower aggregation units, such as regions or cities (e.g. Ezcurra and Rios, 2019; Giannakis and Bruggeman, 2017; Martin, 2012; Östh, Reggiani and Nijkamp, 2018): wherein stands for resilience in city i, is the employment rate in city I at time period t, is the employment rate in city I during the initial period of the analysis t-1, while stands for the employment rate at the national level in year t and is the national level of employment rate in the initial time period t-1.
For be er capturing the capacity of cities to absorb and recover from shocks, our study considers two distinct phases of resilience, namely resistance and recoverability (Martin, 2012). Resistance is computed as the diff erence between maximum employment level before the crisis (2008) and minimum employment level during the crisis (2011), while recovery is computed as the diff erence between the most recent data available a er the crisis (2017) and the minimum employment level during the crisis (2011).  e interpretation is similar for both resistance and recovery. A positive value index means that city i exhibits greater resistance/recoverability to recessionary shock than the national average, while a negative value implies that city i is less resistant or has a lower recoverability than the national average.  e following cross-sectional model is used for estimating the role of creative industries regarding resilience: wherein stands for resilience of city I, while the model is estimated for each of the two distinct phases of resilience, namely resistance and recovery. is the main interest explanatory variable and measures creative industries using diff erent proxies (i.e. such as the share of fi rms and employees activating in creative industries, but also the share of turnover owned by the creative industries). What is more, creative industries were split into 12 distinctive branches in other to allow for an asymmetrical impact across them (please see Table A1 in Appendix A for more details). Finally, despite the variable measuring the creative industries, includes a set of explanatory variables (see Table A2 in Appendix A), includes the unobserved regional (NUTS2) specifi c eff ects and is the error term.  e models were estimated by means of robust OLS, where the spatial dependency tests did not suggest a diff erent approach. In order to avoid reverse causality, the explanatory variables are computed as average during the 2006-2008 period for the re-sistance model, while for the recovery model averages for the 2009-2011 period were used.

Urban resilience in Romania
 e costs of the Great Recession at the end of 2000s have induced a high degree of spatial heterogeneity across the EU regions (Capello, Caragliu and Fratesi, 2016). Romania was no exception, as studies point out diff erent regional reactions to the latest global economic shocks (Zaman and Goschin, 2015; Benedek and Lembcke, 2017), which focuses the research a ention on empirical analysis of urban resilience.
As already described, our study relies on a widely used defi nition of resilience, that is the Martin (2012) sensitivity index.  is approach is more dynamic and corresponds rather to the ecological resilience which 'assumes that systems are characterized by multiple stability domains and that if a shock pushes a system beyond its 'elasticity threshold', the system may move to a diff erent domain or state' (Martin, 2012, p. 7). Nevertheless, this approach assumes that regions are expected to follow the same trajectory of the national economy, and not an autonomous path or return to the pre-shock growth path.  us, we have delimited the two distinct resilience stages by looking at the national evolution of the employment rate ( Figure 2). Given the yearly radiography of employment evolution in Romania, the 2008-2011 period can be regarded as the resistance interval and it covers the change from peak to trough.  e recovery phase is marked by the 2011-2017 interval, namely the timeframe between the minimum during the crisis (trough) and the most recent data available.  Our decision to rely on the labor market data to measure the impact of the Great Recession derives from both theoretical and methodological considerations. In theoretical terms, labor market adjustments may be among the main options available for fi rms in order to reduce costs in times of economic downturns, which make labor market indicators, such as employment, a good predictor for economic fl uctuations (Fingleton, Garretsen and Martin, 2012).  e employment rate was largely used for measuring resilience at lower geographical levels (Simmie and Martin, 2010; Giannakis and Bruggeman, 2017; Kitsos and Bishop, 2018; Östh, Reggiani and Nijkamp, 2018; Ezcurra and Rios, 2019). In methodological terms, other output measures at subnational levels, such as GVA, have been criticized (Gripaios and Bishop, 2006) and, also, are not available at city level. Nevertheless, we need to point out that the present study only relies on the number of employees in the private sector. However, this does not reduce the value of our results, as the private sector employment is much more sensitive to the economic cycles. Furthermore, employment in the public sector was drastically limited in 2009, which smoothed fl uctuations during the recession 2 . Considering the values displayed during the resistance and recovery periods, the following city types were defi ned (a er Martin et al., 2016): a) cities with good resistance and good recoverability; cities with good resistance, but weak recoverability; c) cities with weak resistance, but good recoverability and d) cities with weak resistance and weak recoverability.  erefore, Figure 3 reveals a high heterogeneity in terms of resilience among the Romanian cities. Some cities showed a be er resistance to the crisis, such as Orăștie and Sighetu Marmației, while others were more aff ected, namely Câmpia Turzii, Caracal, Gherla, Codlea and Pașcani. Some of these cities managed to recover rather fast, such as Pașcani and Codlea, but others displayed both poor resistance and recovery, such as Câmpia Turzii, Vaslui, Calafat and Onești.  e large Romanian cities were also aff ected, only Cluj-Napoca, Sibiu, Brașov and București displaying positive values for both resistance and recovery capacities. Given the very diff erent reactions of Romanian cities in the a ermath of the Great Recession, our study seeks to shed more light on the role played by the creative economy.  erefore, Figures 4a-f provide a framework for the relation between the share of creative industries (in terms of enterprises, employees and turnover) and resilience. All fi gures appear to suggest a positive correlation, which clearly hints at a signifi cant role of creative industries related to both city resistance and recovery capacity. However, such conclusion should be made with great caution, since it might be a spurious correlation resulted from the omission of other important variables infl uencing the labor market performance. Previous studies which focused on Romanian cities pointed out more factors aff ecting their resilience capacity. Migration, reducing number of employees and aging were shown to be the main causes behind the so-called 'shrinking cities' (Bănică, Istrate and Muntele, 2017).  ese factors reduced their resilience capacity, as well as accentuating urban decline. Bucharest remarks itself with a higher resilience capacity than the other cities. Despite its spatial and environmental vulnerabilities, the Romanian capital benefi ts from higher social and economic endowments, be er learning capacities and innovative potential .  is comes in line with previous fi ndings, as creative industries were shown to be an important triggering factor of innovative capacity in modern cities, as well as for social and economic growth (Kourtit and Nijkamp, 2013; Cunningham and Po s, 2015). Although focusing on larger cities in Romania, the geography of creative industries reveals a heterogeneous dispersion, both before and during the economic crisis. Using the six criteria displayed in Table 1, the cluster analysis has market 4 clusters in terms of creative industry concentration, with the percentage varying from approximately 1 percent to 9 percent in terms of fi rms.
Overall, the numbers reveal that, on average, creative industries did not recorded decline during the crisis. On the contrary, a slight increase is displayed for all three measurements, as compared to their level in 2008. At the other end of the sample lies the fourth cluster, which includes 23 cities with very low concentration of creative industries. While the creative industries account, on average, less than 4 percent of fi rms, their share in terms of both employee and turnover is below 1 percent.

Creative economy and resilience: estimation results
A s previously discussed, the paper explores the role played by creative industries in the resilience of Romanian cities during the Great Recession by relying on an estimation strategy using ordinary least squares (OLS). As mentioned, a spatial autocorrelation test using Moran I statistic for resilience and recovery variables, as well as for the model errors, did not indicate a signifi cant presence of spatial dependence. Two distinct stages of resilience were considered, namely resistance and recovery (based on Martin et al., 2016). Table B1 in Appendix B displays results for the model explaining the resistance of cities to the economic shock, while Table B2 (Appendix B) resumes results for the recovery model. Both tables have similar structures. While models (1) -(6) display results when the creative industries are measured as a whole, models (7) -(12) account for 12 diff erent creative classes, in order to allow for a diff erent impact across diff erent creative classes.
 e estimation results reveal that larger shares of creative industries are associated with greater regional resistance (model (2) in Table B1), which confi rm our fi rst hypothesis (H1). However, the result is only confi rmed when creative industries were measured as the share of employees (but not when the creative sector was measured by the share of fi rms or the share of turnover). Dropping some of the insignifi cant explanatory variables does not signifi cantly change our previous results (model (4) in Table B1).
Further, models (7) - (12) in Table B1 deepen the analysis of the role of creative industries to cities' resistance.  us, the creative activities were divided in 12 diff erent groups in order to account for a possible asymmetrical impact across them.  e results indicate that cities with larger shares of activities belonging to groups 1 and 2 (Advertising and Architecture) managed to surpass easier the shocks caused by the Great Recession. Larger shares of groups 4 and 6 (Libraries, museums, cultural heritage and Film, video, music, TV, radio) also appear to be positively correlated with greater resistance, but these become statistically signifi cant only a er dropping some of the insignifi cant variables (models (11) and (12) in Table B1). In comparative terms, there are the groups 2 and 4 which show the highest support to cities' resistance, as displayed by the estimated coeffi cients. Given the asymmetrical impact across diff erent creative industries groups, the third hypothesis is also confi rmed (H3).
Among the diff erent groups of creative industries, group 12 (Research activities) is the only one indicating a signifi cant negative sign.  us, cities with larger research activities before the crisis were more aff ected and thus displayed higher employment downturns.
Focusing on Table B2, which reports the results for the model explaining the recovery of cities, we conclude that, unlike the model referring to resistance, the creative industries, measured as a whole, do not seem to explain the recovery of cities in the aftermath of the Great Recession (Models (1) -(3) in Table B2), which rejects our second hypothesis (H2).  e results remain roughly the same when insignifi cant variables are dropped (Models (4) -(6) in Table B2). Nevertheless, when spli ing them into more groups, some asymmetrical infl uences are revealed. On the one hand, t he recovery of cities is positively aff ected by large shares of creative groups 5 and 9 (Design and Publishing -models (7) - (12) in Table B2). On the other hand, higher intensity of creative groups 1 and 11 (Advertising; Cultural education) are associated with slower recovery over the 2011-2017 period (the negative infl uence of creative group 11 is only confi rmed a er insignifi cant variables are dropped -Model (11) in Table 2B). Unlike class 11, one explanation for the negative impact of class 1 might be that this group was shown to support be er resistance to shocks and, thus, a shorter gap to be recovered compared to the pre-crisis level.
Besides creative industries, other factors were also validated as signifi cant for explaining the capacity of cities to resist and recover from shocks. Structural diff erences of the cities' economic activities were confi rmed to ma er for urban resilience. If higher shares of the construction sector translate into lower resistance, higher shares of employment in agriculture and fi nancial intermediation and insurance seems to be related to faster post-crisis recoveries. One explanation for the lower resistance of cities with larger construction sectors might be related to the fact that this sector was one of the triggers of the Great Recessions and, consequently, one of the most aff ected. As for the positive contribution of agriculture, this may be related to the high degree of protection that characterizes agricultural markets in the EU, which helped cities to face easier the shocks and recover faster, returning to pre-crisis employment levels (Rodríguez-Pose and Fratesi, 2007).
An easier access to healthcare services, higher public local investments and a higher degree of decentralization of local budgets appear to be important factors in explaining the be er capacity to resist to economic shocks. On the other hand, higher shares of green spaces, as a proxy for local amenities, and higher share of students appear to negatively relate to the city's capacity to cope with economic shocks.  e negative correlation between the share of students and the resistance capacity may be related to the fact that the young population was among the most vulnerable demographic categories during the Great Recession. Finally, when it comes to recovery, higher income levels per capita support faster recoveries, while higher age dependency levels undermine it.

Conclusions
 is article examines the relation between creative economy and urban resilience in Romania during the Great Recession.  e country is an interesting case study because the impact varied signifi cantly across Romanian cities.  us, we used the rich data set from the Romanian National Trade Register Offi ce to separate 12 distinctive creative groups while testing their importance for both urban resistance and recovery a er the last economic crisis. In addition, other factors stemming from resilience and economic growth studies were also tested.
 e added value of this paper is twofold. First, the study deepens the understanding of the creative economy on national grounds, identifying the creative industries at urban level and assessing their contribution to development, considering employment and turnovers. Second, the research contributes to a topic still not enough tackled in the literature -the potential infl uence of the creative industries on urban resilience, by providing empirical clues derived from the Romanian case.  e estimation results reveal that, when creative industries were accounted as a whole, they were confi rmed as a signifi cant driver for urban resistance, but not to urban recovery.  e division of the creative industries into 12 diff erent groups unveiled an asymmetrical impact across them. If creative groups 1, 2, 4 and 6 (Advertising; Architecture; Libraries, museums, cultural heritage; Film, video, music, TV, radio) appear to be positively correlated with be er resistance, higher shares of group 12 undermine urban resistance. When it comes to recovery, whereas larger shares of creative groups 5 and 9 (Design and Publishing) enhance urban recovery, cities with larger shares of groups 1 and 11 (Advertising; Cultural education) display slower recoveries in the a ermath of crisis.
Furthermore, other factors were also shown to have a signifi cant explanatory power. A be er access to healthcare services, higher local investments and a higher decentralization of local budgets appear to enhance the resistance of cities against economic shocks, whilst higher shares of agriculture and fi nance, as well as a higher income per capita help to a be er urban recovery. At the other end lies the construction sector, local amenities and the share of students which seem to negatively relate to resistance of cities. In the wake of shocks, urban recovery is marred by higher age dependency rates.
 e results of this analysis suggest possible implications for policy initiatives that aim to identify triggers which may enhance urban resilience. Our paper provides indications that rising a ractiveness for the creative economy may act as a buff er to potential economic downturns. In the same time, developing specifi c instruments (e.g. fi nancing funds, learning and training programs, creative hubs initiatives) focused on key creative industries could compensate and sustain them in the recovery process. In order to do so, we consider that an important step further is defi ning a national strategy concerning the creative economy. A nationwide framework for addressing the creative sector is for a more comprehensive and homogenous look on the dimension, nature and domestic specifi cities of the sector (including disparities and threats). Although several researches were conducted on various themes related to the Romanian creative industries, defi ning the economic activities (using NACE codes) included in the sector varies, making the fi ndings and possible implications for the local agendas harder to implement and to integrate at strategic level. Also, it makes the comparison between the fi ndings of various researches more diffi cult.
When assessing the results of the present study, another concern relates to the limited perspective on development used when analyzing the economic resilience of the cities. Relying solely on the share of enterprises, turnover and employees does not provide insights about the creative sectors' relation with other components of development, or about their possible spillover eff ects which may infl uence resilience a er a crisis, even on lower recovery rates.
 erefore, our study is a preliminary step towards a be er understanding of the connection between creative economy and urban resilience. While off ering reasons for a closer and deeper look at the creative economy when designing strategies seeking to enhance the resistance and recovery capacities of the cities, our work initiated a fresh approach on the role played by creative industries in urban resilience. Future research could try capturing the whole creative industries by including both private and public sectors. At the same time, a diff erent methodological se ing that accounts for a potential simultaneity between creative industries and resilience might also strengthen our fi ndings. Considering the recent changes in the creative class' movement pa erns, our analysis could be applied on a wider geographical area, including both urban and rural se ings. Comerț cu amănuntul al discurilor şi benzilor magnetice cu sau fără înregistrări audio/video, în magazine specializate 5911 Activități de producție cinematografi că, video şi de programe de televiziune 5912 Activități post-producție cinematografi că, video şi de programe de televiziune 5913 Activități de distribuție a fi lmelor cinematografi ce, video şi a programelor de televiziune 5914 Proiecția de fi lme cinematografi ce 5920 Activități de realizare a înregistrărilor audio şi activități de editare muzicala 6010 Activități de difuzare a programelor de radio 6020 Activități de difuzare a programelor de televiziune 6391 Activități ale agențiilor de știri 7722 Închirierea de casete video şi discuri (CD-uri, DVD-uri)

Appendix A
Recording reproduction Fabrication of magnetic and optical media for recording Fabrication of musical instruments Retail sale of magnetic discs and tapes with or without audio/video recordings, in specialized stores Film, video and television program production activities Post-production activities of fi lm, video and television programs Film, video and television program distribution activities Movies projection Audio recording and music editing activities Radio broadcasting activities Television broadcasting activities Activities of news agencies Rental of video cassettes and discs (CDs, DVDs)        Notes: Robust standard errors are given in parentheses. Signifi cance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. The dependent variable was computed as a relative change to the national level over the period 2011-2017. Explanatory variables are averaged during the 2009-2011 period, except for initial employment level which refers to 2011.