Be Water my Friend: Building a Liquid Destination through Collaborative Networks

11 12

 Planning to change and broadening participation increase organisations' ability to capitalise on opportunities that arise in an uncertainty situation;  DMOs must stimulate organizations' preparedness and their participation in the destination's planning process.

Introduction 19
Measuring a tourism destination's resilience level is not easy. The task represents a real 20 challenge for managers, as it requires them to evaluate economic and social systems in a 21 context of change and high uncertainty, since tourism in embedded in human actions. 22 Responding adequately to uncertainty is key to building efficient adaptation capabilities (Cheer 23 & Lew, 2017). In this context, resilient destinations are able to cope with sudden changes 24 (Luthe & Wyss, 2014). The concept of resiliente has been extensively discussed in studies on 25 the field of psychology, as well as in post-disaster studies. In the context of tourism 26 destinations and organisations, however, it has not yet received enough academic attention, 27 therefore, there are still some knowledge gaps to be filled. 28 In recent years, important advances have been made towards measuring tourism 29 organizations' and destinations' resilience. In this context, several studies have proposed 30 models and frameworks resulting from qualitative (Becken, 2013) and quantitative approaches 31 (Brown,  resilient, according to which one of the distinctive elements of resilience is a system's ability to 34 absorb disturbances and evolve, especially by self-organising without resorting to external 35 drivers (Prayag, 2018;Walker, Holling, Carpenter, & Kinzing, 2004). In a resilient system, this 36 endogenous self-organization happens both in extraordinary circumstances of change (e.g., 37 great disasters) and in the face of incremental changes (Lew, 2014). Therefore, resilient 38 organisations face crises as sources of strategic opportunities, as they find opportunities in 39 times of distress. 40 Among the factors that contribute to the self-organisation ability, and therefore, to a 41 company's resilience, is the development of social networks and community engagement 42 (Biggs, Schlüter, & Schoon;Cutter, Burton, & Emrich, 2010). The former provides a 43 different approach in the study of crisis management and resilience, with an emphasis on the 44 interactions between organizations (Scott & Laws, 2005). It is critical that policy makers and 45 other stakeholders (e.g., government, civil society, academia, and the media) work together to 46 2 create more agile and adaptable forms of local, national, and global governance and risk 47 management (WEF, 2017). 48 Considering the extant literature on organisational resilience, especially in the context of 49 tourism organisations, the present study builds on several specific contributions. First, it has 50 been considered that tourism organisation managers' view about their crisis preparedness and 51 their participation in the destination's crisis plans (which are normally led and coordinated by 52 the DMOs) may foster the organisation's resilience. This approach builds on previous studies' 53 results (Biggs,  reducing disruptions, as they lead to mutual learning via shared experiences. In this context, a 56 key-concept in the present study is that collaboration between companies, as well as regular 57 interactions between public and private organisations lead to an agile and flexible (or 58 adaptable) environment (Islam & Walkerden, 2015;Getz & Timur, 2005). It has also been 59 considered that, in a context of crisis, enhancements in resilience are measured by the 60 company's ability to seize opportunities that may arise (Seville et al., 2006). There is also a 61 wide theoretical basis that backs this approach, which will be minutely reviewed in the 62 following section. 63 The factors mentioned above have been recognised as enhancers of resilience, as they are 64 related to organisations' behaviour in the pursuit of proactive resilience building. Within the 65 context of tourism destinations and organisations, however, no study has empirically assessed 66 the role of the factors that contribute to resilience building. In this context the present paper 67 aimed at assessing the role of two of the main factors that contribute to the development of 68 resilience in tourism organisations, more specifically, hotels: their preparedness to crisis, and 69 their participation in the destination's crisis management plans. 70 To achieve the mentioned goal, a quantitative survey was carried out with hotel managers in a 71 destination particularly affected by two types of natural disasters: forest fires and oil spills. 72 Hotels have been chosen as the object of research because, along with DMOs, they play a 73 crucial role in achieving inter-stakeholder collaboration for sustainable destination 74 development, as different actors within the industry trust or depend on them (Timur & Getz,75 2008; Sheehan & Ritchie, 2005). 76 The collected data has been analysed through logistic regression models. Results show that 77 both variables -preparedness and participation -play an important role in building an 78 organisations' ability to adapt and seize opportunities that might arise in times of change or 79 disturbing events. Therefore, the study's findings demonstrate that tourism organisations' 80 preparedness and participation in a destination's crisis management plans indeed contribute 81 to enhance their resilience, and consequently, the destination's.  Kierkegaard's will to meaning, the author analysed how individuals react in face of adversity by  94  examining people who had suffered from the experience of the concentration camps during  95 World War II, himself included (Frankl, 1985). Later, Rutter (1990) defined resilience as an 96 individual's positive reaction to stress and adversity; Masten (1989) defined it as the positive  97  part of an adaptation following extenuating circumstances; and Garmezy (1991)  98 conceptualised it as the capacity of recuperation and maintenance of adaptive functioning in 99 face of disability. In this context, resilience became a common concept in the field of 100 psychology (Ryff et al., 1998). Accordingly, within the context of psychopathologic 101 development, resilience refers to a positive adaptation to adversity (Luthar, 2015). 102 The concept has been extrapolated to other disciplines and became a popular term (Meerow,103 Newell resilience, which is strengthened by certain personal characteristics (Frankl, 1985), the 110 resilience of tourism destinations and organisations is also favoured by specific organisational 111 factors. Amongst those, are a company's self-organisation ability, its social networking, and its 112 stakeholders' participation in the destinations' crisis management plans. 113

Self-organisation ability as a distinctive element of resilience 114
Resilience refers to one's ability to respond to change (Holling, 1973). Learning and adaptive 115 change allow systems to successfully respond, recover and adapt to new conditions. Adaptive 116 resilience includes social learning by individuals, governance structures, or stakeholders after a 117 triggering event (Cutter, 2016). In this context, resilience acquires a positive social meaning, as 118 it is preferable to employ the concept of adaptation rather than other, such as vulnerability 119 (Weichselgartner & Kelman, 2014). Disturbances in daily activities make it possible to observe 120 how a system works in such situations, namely, how it responds to the changes and where its 121 limits lay. This allows managers to identify potential failures, and consequently, improve the 122 company's ability to adapt (Woods, 2017). 123 Resilient structures present a high level of internal self-organisation, which is an endogenous 124 process, as it cannot be simply forced by exogenous drivers (Prayag, 2018;Walker, Holling, 125 Carpenter, & Kinzing, 2004). Such ability to re-construct is more than a simple passive 126 resistance to negative aspects followed by a mere adaptation to the environment 127 (mithridatism). Instead, a resilient system develops active resistance to external impacts that 128 can affect its normal functioning. Regarding socio-ecological systems, although the system re-129 organises spontaneously, human actors' abilities and interests significantly affect the system's 130 recovery capabilities (Walker et al., 2004). 131 According to Ruiz-Ballesteros (2011), prompting the opportunity of self-organization is an 132 enhancer of resilience in tourism development. changes, and the solidary cohesion among partners (Getz & Timur, 2005). 170 Collaboration between companies also plays an important role in reducing disruptions, as it 171 leads to mutual learning via shared experiences. This allows each business to operate more 172 safely and coordinate prevention efforts with their partners. Ultimately, this results in a 173 significant reduction of many potential disruptions that originate outside the organisations or 174 spread to others (Sheffi, 2005). In contrast, within the traditional firm-centred approach to 175 strategy, social networking managers need to formulate their strategies in terms of the 176 collective success of their networks, as well as their individual organisations' self-interest 177 (Bresser, 1988). In this context, as indicated by several studies (Lee, Vargo, & Seville, 2013; 178 Sautter & Leisen, 1999), coordination between different organizations can facilitate both 179 disaster preparedness -planned resilience -and response to situations that had not been 180 considered during planning -adaptive resilience. 181

Stakeholders' participation in the destination's crisis management plan 182
According to the stakeholder theory, an organisation maintains relationships with many groups 183 of individuals, including employers, clients, suppliers, governments, and local communities 184 (Freeman, 2010). The development of a destination crisis management plan is the 185 responsibility of DMOs, which must also foster the participation of all local tourism 186 stakeholders (Paraskevas & Arendell, 2007). In the present work, it is considered that tourism 187 stakeholders' participation in the destination's crisis management plan may vary from the 188 mere passive attendance of public hearings and advisory committees to other forms that imply 189 5 a greater involvement in the destination's decision-making process (Byrd, 2007). To carry out 190 the destinations' crisis management plan, DMOs must achieve a high level of agreement and 191 participation amongst tourism stakeholders (Pauchant & Mitroff, 1992). In this context, 192 tourism organisations and DMOs can benefit from working together in order to be more 193 prepared to face the effects of crises, as well as to overcome potential risks (d' Angella & Go,194 2009; Johnson, Lu, Tolomiczenko, & Gellatly, 2008). 195 The participation of many stakeholders in recovery plans and in decision making processes is a 196 key factor for an efficient network coordination (WTO, 2000; Jamal & Getz, 1995). In tourism 197 destinations, resilient thinking can be implemented through the strengthening of relationships 198 and cooperation, as well as through the construction of a comprehensive social network. All of 199 these factors aim at allowing a diverse group of stakeholders (e.g., business managers, 200 employees, local government, residents, and tourism planners) to participate in the planning 201 and decision-making processes (Zacher, 2018;Sautter & Leisen, 1999). preparedness improves knowledge on the organisation itself, which generates competitive 217 advantages for destinations. This allows for a quick identification of unfavourable situations' 218 effects on the organisation (Glaesser, 2003). Planning for crisis -and its integration in strategic 219 planning -leads to knowledge on products and markets, which builds organisations' ability 220 react faster, and thus, limit the impacts of disturbances in their businesses (Johnson et  problems, allow managers to identify personal that must be replaced, favour the development 229 of new strategies and market segments, enhance early alert systems, and help accelerating 230 changes (Burnett, 1998). In resilience terms, preparation and planning are effective because 231 they enhance awareness of critical dependencies and functions within the organisation. 232 Moreover, they provide confidence to seek out opportunities even in times of disturbance 233 (Seville et al., 2006). In this context, resilient organisations have the foresight and situation 234 awareness necessary to prevent the emergence of potential crises (Holbeche, 2015). 235 A high level of participation in the destinations' crisis management plans also increases trust 236 and acceptance of tourism plans, and consequently, their political legitimation (Bramwell &  237  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  organisation's ability to adapt and seize opportunities in times of disturbances. 263 Managers' perceptions on these characteristics were measured through a 7-point Likert scale 264 (1: strongly disagree; 7: strongly agree), through which they identified the roles that best 265 reflected their behaviour. When operationalising a categorical variable, the measure of 266 gradualness is particularly important. In this context, preparedness was measured by the item 267 "The preparation and planning for the crisis of my organization is adequate". Before evaluating 268 their preparedness to crisis, however, respondents had to answer a block of seven qualitative 269 dichotomous questions, which represented elements that indicate good preparedness. This 270 aimed at ensuring that respondents were familiar with the concept. Moreover, it has been 271 considered that a crisis results from an unforeseen event over which one has little or no 272 control (Glaesser, 2003). Such unpredictability extenuates the need for preparedness for 273 different scenarios and the threats they bring about. Therefore, the aspects that represented 274 preparedness in the questionnaire were generic and applicable to any situation. 275 Participation, in turn, was measured through the item "My organization's participation in the 276 destinations' crisis management plans was very high". Hotels can participate in a destiantions's 277 crisis management plans in multiple formal and informal ways (Byrd, 2007). In this context, the 278 term "very high" was employed to encompass any of these possible forms of hotels 279 involvement, as it refers to the intensity of the participation, rather than a to particular form. 280 Moreover, it aimed to mitigate possible overvaluations of the hotel's participation by 281 respondents (Hoorens, 1995). 282 The dependent variable Opportunities seized from the crisis comprised four items: Opp1, 283 Opp2, Opp3, and Opp4, consisting of numeric variables and operationalised through the 284 mentioned 7-point Likert scale. Resilient organisations can turn crises into a source of strategic 285  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65   7 opportunities (Vargo & Seville, 2011). Therefore, they find opportunity in times of distress 286 (Stephenson et al., 2010). The present investigation faced the difficulty in measuring strategic 287 opportunities to organisations. In the absence of previously validated measurement 288 instruments, an ex novo questionnaire was developed. The literature review points to four 289 parameters of opportunities that may be presented to tourism businesses during crises or 290 disruptive events. These items are described in Table 1. 291 An initial univariate analysis was carried out to determine which independent variables are 307 significantly related to the dependent variables. To this end, the probit function was employed 308 as the link function for the parameter estimation for the dependent variables. Preparedness 309 and participation scored significantly (p < .05) in all the contrasts, with all positive and 310 significant coefficients in the Wald statistic test (see Table 2). 311 In section 3.3, which presents the analysis methods and the model proposal, the correlation 313 coefficient between the two independent variables is tested. This aimed at ensuring that the 314 inclusion of both these variables in the model did not cause multicollinearity, as they are 315 linearly related. 316

Data collection procedures 317
The study was carried out in the autonomous community of Galicia, in the northwest of Spain. 318 This particular region was chosen because it regularly suffers two major types of disasters: oil 319 spills, caused by oil tankers sinking near the coast, and forest fires. The data collection process comprised four different steps. First, each hotel manager was 339 contacted via telephone by a member of the research team, who explained the objectives of 340 the study and solicited their participation. From the 226 3 to 5-star hotels in the region, 221 341 were contacted, and only one immediately refused to participate. The second step consisted in 342 sending an e-mail to the director or the manager of each hotel, which aimed at reinforcing the 343 commitment previously established via telephone. The third step consisted in sending a postal 344 package including an explanatory letter, the research questionnaire, and a prepaid envelope 345 with the researcher's address to facilitate the task of sending back the filled in questionnaire. 346 Finally, in the fourth step, another e-mail was sent to the hotel representatives who had 347 agreed to participate in the research but had not sent back the filled in questionnaire. 348 Throughout all those steps, some measures were taken in order to avoid non-response bias, 349 such as ensuring data confidentiality (Cruz Cantero, 1986) and the data collection procedures' 350 homogeneity (Azorín & Sánchez-Crespo, 1986). In the end, a total 72 valid questionnaires were 351 obtained, which implies a no-response rate of 67%. The non-response rate does not 352 necessarily denote a lack of interest in the subject, as natural disasters have significantly 353 affected tourist arrivals and hotel occupancy rates in Galicia (Loureiro & Barrio, 2009), as well 354 as received intense media coverage. Therefore, the authors infer that it rather indicates a 355 certain resistance to surveys by hotel managers, as previously pointed out by Goyder (1985). In 356 the case of Galician hotels, this is likely intensified by their obligation to respond the local 357 tourism authority's hotel occupancy survey. In order to attenuate the non-response caused by 358 this context, the researchers attempted to demonstrate commitment by contacting managers 359 9 in different phases, but also avoided being intrusive. Despite these efforts, the non-response 360 rate was still quite high. Nevertheless, due to the addressed circumstances, the there is no 361 reason to believe that non-response bias is an issue within the present study. 362

Data analysis and model proposal 363
The collected data has been analysed through ordinal logistic regression, which allows for 364 measuring the effect of a covariable matrix's value on the cumulative probability of a particular 365 event (in the present case, the dependent variable "Opportunities to the organisation") taking 366 place. The use of logistic regression analysis in tourism research has grown significantly in 367 recent years. Between 1998 and 2002, logistic regressions are employed in 3.06% of tourism 368 studies (Palmer, Sese, & Montano, 2004). Depending on the study's design and data collection 369 procedures, logistic regression may serve one or more of the following objectives: 1) 370 determining the existence of a relationship or association between the independent variables 371 X i and the dependent variables Y; 2) measuring the magnitude of such relationship; and 3) 372 estimating the probability of a certain event happening as an outcome of the independent 373 variables' X i values (Jovell, 2006). 374 In ordinal logistic regression, the dependent variable Y is expressed through a categorical 375 ordinal value, e.g., 1, 2, … J, which is affected by a vector of the explicative covariables' values, 376 that is: X = (x 1 , x 2 , ..., x n ). The cumulative probability of an event is limited to a certain value, 377 which represents an advantage of logistic regression analysis when compared to other analysis 378 techniques (Winship & Mare, 1984). The starting equation in ordinal logistic regression models 379 is based on the following expression: 380 The ratio between the probability of Y taking a lower value and that of it taking a higher value 382 is the odds ratio. Such coefficient is calculated through the following expression: 383 1, 2 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  contrasted through a test of parallel lines. 423

Sample Characterisation 425
Respondents were mainly hotel directors and general managers (80.6%), department 426 managers (6.9%), and owners (5.6%). The sample consists mostly of 3-star (59.7% of the 427 sample) and 4-star hotels (33.3%). Although only five 5-star hotels responded (6.9% of the 428 sample), they represent 62.5% of all high-class hotels in Galicia. The distribution of responses 429 throughout the Galician provinces, as well as in terms of hotel category, is proportional to the 430 total population's distribution, which reinforces the absence of non-response bias. 36.1% of 431 hotels specifically target leisure travellers, while 18.1% prioritise business travellers. The 432 remaining 41.7% work with both markets. To evaluate hotels' perceived level of risk, a 7-point 433 scale (1: without risk; 7: very high risk) was employed. The highest perceived levels of risk refer 434 to fires (5.86) and oil spills. Negative perceptions of residents towards tourism, on the other 435 hand, had the lowest rated perceived level of risk (1.74). 436

Logistic regression model 437
The data was analysed though IBM SPSS v.17.0. Results are summarised in Table 3. 438 to the exploratory character of the present study, the relationships between the dependent 471 variable and the covariables must be interpreted in terms of association or correlation, rather 472 than considering an evidence of predictive effectiveness. Logistic regression models applied to 473 a transversal design estimate events that took place in a specific time period, in which the 474 independent and the dependent variables to be included in the model could also be observed. 475 Therefore, the results they yield cannot be extrapolated to future or hypothetical situations. 476 In the context of the models presented in this study, the variables preparedness and 477 participation are correlated to a set of opportunities that emerge to hotels in crisis or change 478 situations. The negative sign in the β j coefficients implies that higher coefficients indicate a 479 decrease in the odds ratio, which, is the ratio between the likeliness that Y takes lower values 480 and that of it taking higher values. All coefficients obtained are positive. Therefore, an increase 481 in the hotels' preparedness and participation enhances the likeliness of them seizing a set of 482 opportunities. 483 Disasters are characterised by their chaotic and unpredictable nature (Hosie y Pffor, 2009 Moreover, preparedness contributes to reactivating the company's strategic planning, and 497 consequently, to acquiring knowledge of the product and its markets, identifying staff 498 members in need of renovation, and developing new strategies. 499 Galicia's specific characteristics provide certain particularities to the study. The region has 500 been afflicted with forest fires and oil spills. These events do not catch hotel managers by 501 surprise, as they take place with relative frequency. In fact, these two types of disaster showed 502 to be the events with highest perceived risk among the respondents. Such experience is a good 503 basis for improvements in learning and knowledge transfer. On the other hand, the local 504 population seemingly has very good perceptions regarding tourism, as Negative perceptions of 505 residents towards tourism is the item with lowest perceived risk level. This characteristic, along 506 with visitors' positive evaluation on local hospitality, strongly influences support for tourism 507 development and participation in the planning and management processes (Rasoolimanesh,508 Jaafar, Kock, & Ramayah, 2015). 509 A clear implication of the present study's results is the confirmation that stimulating hotels' 510 preparedness and participation in the destination's planning process contributes to their 511 resilient capacity. Each hotel has the responsibility to undertake a strategic planning that 512 accounts for unforeseen and undesired situations. However, promotion and stimulation 513 actions must be carried out by organisations with that have the necessary resources. The 514 active and constant commitment with resilience thinking implementation in a tourism 515 destination may be a difficult task to the entity responsible (Zacher, 2018). Therefore, such 516 responsibility should be attributed to entities that possess the necessary competencies to fulfil 517 it. Many authors agree that such task is the responsibility of Destination Management 518 Organisations (DMOs), as they play a central role in setting up the destination (Zacher, 2018;  519  13 Luthe & Wyss, 2016). The activities developed by the DMO, such as marketing planning and 520 development, are directly connected to resilience at the destination level. The main function 521 of a DMO is to coordinate stakeholders towards a common vision. The ability to carry out this 522 task is determined by the quantity and the quality of contacts the DMO has with local 523 stakeholders and the governmental agencies (Paraskevas & Arendell, 2007). In this context, 524 according to Ritchie and Crouch (2003), DMOs play a critical role in ensuring that stakeholders' 525 expectations are met. 526 Organizations' collaboration in social networks allows them to adapt to changes more easily, 527 since one of the problems that are typically detected in catastrophe situations is precisely the 528 lack of communication between stakeholders, along with insufficient planning for disaster 529 management (Mair, Ritchie, & Walters, 2016). In this context, the ability to make network 530 connections and work with organisations in the community adds to a hotel's resilience to 531 disasters (Orchiston, Prayag, & Brown, 2016). The presence of social networks contributes to 532 organisations' agility and innovation. Moreover, it allows them to deal more fluidly with their 533 own complexity (Holbeche, 2015). Nevertheless, social networks still have a considerable 534 expansion potential (Brown et al., 2019). Therefore, their role in building organisations' 535 resilience might still be significantly amplified. 536 The theoretical development of the concept of agile, adaptable, and innovating organisations 537 is connected with that of the liquid organisation, which originates from Bauman (2013). The 538 liquid organization arises as a response to solid and stable organizational structures, which are 539 typical of the industrial society (Bauman, 2013). In times of uncertainty, a liquid organization 540 behaves better: it values the importance of functionality, manages to adapt and transform 541 itself as a response to external -and internal -changes, and focuses on the simplicity of 542 processes to improve their flow (Wearing & Hughes, 2014). 543 A destination can be viewed as a complex and adaptable system. Information and knowledge 544 flows are important factors for such system's general wellbeing (Baggio,Scott,& Cooper,545 2010). Rather than stagnant structures fixed in a specific moment, destinations should be 546 viewed as co-evolutionary processes (Inkpen & Currall, 2004). In this context, destinations that 547 foster collaborative networks, and thus facilitate information and knowledge flow, could be 548 characterised as liquid destinations. 549 The concept of liquid destination is a new proposition. Unlike what might intuitively be 550 inferred at a first glance, the term does not refer to a destination that has diluted the bases 551 upon which it once stood. On the contrary, it designates a destination that builds upon these 552 bases -its nuclear resources -to incorporate flexibility as an endogenous element of 553 resilience. Therefore, a liquid destination is indeed a sound destination, which prospers in 554 contexts of change and disturbances by rapidly adapting and overcoming the consequent 555 obstacles. The so-called rebirth areas, or "phoenix tourism" destinations, as New Orleans after 556 Hurricane to their outstanding resilience to disturbances and risks, such destinations managed to not 558 only recover from the disasters that afflicted them, but in doing so, improve their image. In 559 sum, liquid destinations are more proactive, and therefore, have the ability to quickly adapt to 560 different circumstances in order to create more agile forms of local, national and global 561 governance and risk management (WEF, 2017). Ultimately, this makes those destinations more 562 resilient and allows them to constantly evolve. 563

Conclusion 564
The present paper aimed at assessing the role of two of the main factors that contribute to the 565 development of resilience in tourism organisations: their preparedness to crisis, and their 566 participation in the destination's crisis management plan. To this end, quantitative data was 567 14 collected through a survey with hotels from a region frequently affected by disasters: Galicia, 568 Spain. Hotels were chosen as the object of study due to the role they play, along with DMOs, in 569 the construction of tourism stakeholders' social network, which is a starting point to building 570 destination resilience (Timur & Getz, 2008;Sheehan & Ritchie, 2005). 571 The findings, obtained through a logistic regression, provide an original contribution to tourism 572 literature, as they demonstrate that organisations that prepare for disturbances and 573 participate in the destination's crisis management planning process are indeed better able to 574 adapt, improve and seize opportunities in times of crises. In a context of high uncertainty, the 575 crisis planning process contributes to adaptive resilience, as it provides better social learning 576 for the parts involved, and consequently, generates more proactive postures. Potential 577 improvements presented to organisations include: increased swiftness and efficacy in 578 responding to disturbances, mutual learning between stakeholders due to shared experiences, 579 knowledge development and transfer to employees, and improved staff commitment. The 580 study also provides a methodological contribution, as it is the first to apply logistic regression 581 to the context of resilience in tourism destinations. 582 The findings also bring about important implications for tourism management and policy 583 making. Governmental agencies, especially DMOs, must stimulate hotels' preparedness and 584 their participation in the destination's planning process. Involving destination stakeholders in 585 such processes contributes to the effectiveness of tourism development and crisis 586 management plans, as well as reinforces their legitimacy. Therefore, governmental agencies 587 should facilitate stakeholders' participation in different phases of planning, as well as foster 588 the integration of the destination's and the organisations' strategic planning. 589 Future investigations must attempt to provide contributions to the practical application of 590 resilience measures and attitudes. Given the constantly changing environments that 591 characterise complex systems -in which resilient thinking is centred -longitudinal studies are 592 necessary to assess their evolution. Complex systems, such as tourism destinations, require a 593 multi-dimensional approach. Research based exclusively on the examination of quantitative 594 elements does encompass the ample set of meaning related to the concept of resilient 595 thinking, which integrates several epistemologies (Becken, 2013). Therefore, for a better 596 interpretation of how participation in collaborative networks provides agility and fluidity to 597 elements of the tourism system -strategies, stakeholders, and resources -, the results of the 598 present study must be complemented by those of investigations that examine such 599 relationship from a qualitative stance. 600 3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65