Investigating the Motivations and Constraints of Chinese Peer-to-Peer Accommodation Hosts

Purpose – This research aims to identify peer-to-peer (P2P) accommodation hosts’ perceived motivations and constraints, to examine the prediction of the motivation and constraint factors on hosts’ intention to continue business based on hosts’ attitudes, and to explore the moderating role of business scale. Design/methodology/approach – A scale for hosts’ perceived motivators and constraints was developed. Mixed methods were used to develop and analyse a conceptual framework for demonstrating how constraints and motivations influence hosts’ behavioural intentions. Findings from interviews with hosts interpretatively supported the survey results. Findings – Chinese hosts’ perceived constraints and motivators are identified and explained. The survey results indicate that constraints lower intention to continue one’s business and motivators heightens it. Motivators have a higher effect on attitudes and intentions than constraints do. Business scale was confirmed as a moderator in the constraint—attitude link but not in the motivator—attitude relationship. Practical implications – This paper offers policy implications for governments, online platforms and hosts in terms of establishing incentives and solving problems so that Chinese hosts can sustainably operate their businesses. Originality/value – This paper identifies constraints and motivators and develops a measurement scale for both simultaneously, which provides a holistic explanation of hosts’ attitude and behavioural intention. It also reveals the moderating role of business scale. In investigating the thoughts of existing hosts operating on global and local platforms in China, this paper complements the literature, which mainly focuses on the Western context and a single global platform.


Introduction
The sharing economy has grown rapidly over the last decade, mainly due to the recognition of shared values and improvements in technology (OECD, 2016). The global revenues generated by the sharing economy, including hospitality, dining, car sharing, finance, staffing, and music and video streaming, are estimated to reach around US$335 billion by 2025 (PwC, 2015). The sharing economy is based on digital platforms that share underused assets (Prayag and Ozanne, 2018). The tourism and hospitality sector is a dominant participant in and contributor to the sharing economy due to peer-to-peer (P2P) accommodation (Cheng and Edwards, 2019). After one of the earliest P2P accommodation platforms, Airbnb, was launched in 2007, the P2P accommodation business rapidly gained popularity among guests globally (Guttentag, 2015;Tham, 2016).
The sharing economy highlights the value of sharing, and its business activities heavily rely on interactions between hosts and guests. The boom in the P2P industry largely depends on the diversity of host offerings, and the platforms are merely the intermedium. These factors highlight the overlooked importance of hosts in the sharing economy. Compared to P2P accommodation consumers, who have been extensively evaluated in the literature (Karlsson and Dolnicar 2016;Moon et al. 2019), hosts' attitudes and behaviours have received little attention (Stors and Kagermeier, 2017). The few studies that have examined motivations and constraints have either focused on income and social interactions as motivations (Benoit et al. 2017;Farmaki, 2019;Farmaki and Stergiou, 2019;Karlsson and Dolnicar, 2016) or on perceived risks as constraints (Alraeeini et al., 2019;Malazizi et al., 2018). In this type of business, constraints and motivations often co-exist.
The majority of the extant research has emphasised motivation factors and neglected the constraint factors that hosts encounter. These studies have attached importance only to the growth of the sharing-economy market, resulting in a lack of investigation into conflicts with other stakeholders, such as other residents and platforms, and the challenges posed by government regulations regarding safety, quality and hygiene. Thus, there is a gap in the literature concerning how such factors shape hosts' intentions to continue running their P2P accommodation businesses. Previous studies were conducted during the early stages of the P2P sharing-accommodation industry, when it was undergoing dramatic growth. The sharing model has now entered its second decade, during which a crisis has arisen, requiring the exploration of emerging issues. This study identifies and examines the constraints and motivators in a single model to provide a holistic, robust explanation of hosts' attitudes and behavioural intentions.
This study also explores whether the relationships of constraint-attitude and motivatorattitude are contingent on business scale (single versus multiple properties). The implications of these findings can be suggested to hosts operating at different scales.
This research aims to identify hosts' perceived motivations and constraints, develop a measurement scale, and validate the predictive power of motivators and constraints on overall attitudes and behavioural intentions regarding future operation. Prior studies have mostly focused on Western countries and Airbnb, the implications of which may not be generalisable to Asian countries where local platforms have proved more popular than Airbnb. The present research focuses on the Chinese market and offers policy implications for relevant government departments and online platforms for establishing incentives and solving problems so that hosts can operate sustainable businesses.
This study makes four contributions. First, it enriches understanding of the sharing economy by identifying the motivations and constraints perceived by service hosts operating P2P accommodation and by developing a measurement scale. It contributes to the formation of a holistic picture of the sharing economy by complementing previous studies, which mainly focused on motivators. Second, the understanding of P2P accommodation management and experience has been restricted by the overwhelming focus on the Western context. Empirical evidence drawn from Chinese hosts provides implications that are relevant for Chinese destination management. Third, it demonstrates the power of various motivations and constraints in explaining business owners' continuance intention, whereas the literature has tended to examine the intention to join the industry. Fourth, previous research has focused on the predictors of hosts' attitudes towards and intentions to operate P2P accommodation. Less is known about moderating factors. This study explores business scale as a boundary condition that differentiates constraint-attitude and motivator-attitude relationships.

Research on P2P accommodation hosts
The satisfaction level and perceptions of the supply side of the sharing economy are believed to be as crucial as the demand side in influencing the sustainable development of this business (Dolnicar and Talebi, 2020). Many studies have focused on the demographic characteristics of hosts (e.g., profile photos, names and gender on online platforms) and their non-demographic attributes to examine how such factors affect hosts' business performance (Ert and Fleischer, 2019;Xie et al., 2017). 6 The literature has explored hosts' attitudes towards platforms and guests, and some studies have examined hosts' behavioural intentions in relation to accepting/denying bookings and disclosing information. Wang et al. (2019) assessed the antecedents and consequences of hosts' trust in Airbnb and found that technical antecedents, including the quality of the system, service and information, had the strongest impact on hosts' trust in the platform. Karlsson et al. (2017) examined the factors influencing hosts' decision to accept or deny a booking. Their findings showed that hosts decide to accept a booking request based on age, guest self-assessment, length of booking, trip motivation, travel party, number of people in the profile picture and availability of the profile picture. Several studies have evaluated hosts' and guests' perceptions and behaviours, many of which have focused on the interactions between the two key stakeholders. Moon et al. (2019) found that hosts perceived fewer positive overall interactions and appeared to have a lower level of encounter satisfaction and behavioural intention compared with guests.
Airbnb and Couchsurfing are well-known P2P accommodation businesses (Prayag and Ozanne, 2018). Many studies on the sharing economy have been conducted in the context of Western countries and focused on mainstream online platforms, particularly Airbnb (Dolnicar, 2019), at the expense of diversity in the knowledge construction of the sharing economy. With the swift growth of P2P accommodation supply and demand in the Asian market and the launch of diversified local platforms, it is crucial to examine the operations of P2P accommodation in Asia. Studies on Asian countries and non-Western platforms are limited, perhaps due to restrictions on data accessibility. The current research enriches the literature by collecting data in China on hosts' perceived motivations for and constraints on using Airbnb as well as local platforms.

Motivations and constraints of P2P accommodation hosts
Although the importance of research on the hosts of P2P accommodation has been recognised (Prayag and Ozanne, 2018), research remains limited on the motivation and constraint factors influencing hosts' accommodation sharing (Wang et al., 2019). Service providers in the sharing economy run businesses based on different motivations and constraints (Böcker and Meelen, 2017). Karlsson and Dolnicar (2016) summarised three categories of motives: 1) income, which includes paying bills, making money and affording luxuries; 2) social interaction, which means meeting and interacting with guests from various backgrounds; and 3) sharing, which mainly refers to the sharing of space, although several respondents mentioned the sharing of joy. Among the three categories, economic reasons and social interaction have been identified as predominant drivers (e.g., Alrawadieh and Alrawadieh, 2018;Benoit et al., 2017;Farmaki, 2019;Farmaki and Stergiou, 2019). From their interviews with Airbnb hosts and guests, Farmaki and Stergiou (2019) learned that reducing loneliness was a key driver for hosts and guests to participate in a sharing business. Moreover, Benoit et al. (2017) captured entrepreneurial freedom, including the flexible offering and individualization of service as an important motivator for service providers and confirmed the positive role of electronic sharing systems. Based on Airbnb host data collected in Dubai, Alraeeini et al. (2019) discovered the low costs and high value linked to accommodation sharing and having a user-friendly, highly informative accommodation-sharing system as the main drivers of accommodation sharing.
Sustainable and environmental considerations, such as sharing energy and water resources, are also important reasons for participating in the sharing economy (Hamari et al., 2015). These findings may reflect that the fundamental value of this business model is to share.
The main constraint identified in the literature is risk. Malazizi et al. (2018) revealed that two types of risk, financial and safety/security, negatively affect Airbnb hosts' satisfaction and behavioural intentions, along with political risk. Farmaki (2019) pointed out that female hosts may face an increased risk to their safety. This assumption could be based on Airbnb's 'guestfirst approach', which restricts females from rejecting or cancelling bookings about which they may feel uncomfortable. Apart from risks, limited knowledge about platforms and underdeveloped legislation have also constrained hosts (Alraeeini et al., 2019;Malazizi et al., 2018). Alraeeini et al. (2019) showed that limited understanding of the use of new platforms and insufficient or underdeveloped legislation generated doubts over service quality and the benefits/losses of the business. Stiff competition was found to be another key challenge (Alrawadieh and Alrawadieh, 2018).
Research on the motivation and constraint factors of hosts has been scattered. Although Mahadevan (2019) examined motivators and barriers for hosts, she captured perceptions of guests as would-be hosts rather than existing hosts. The findings from that study indicate that a sharing philosophy is a dominant motive. In contrast, opportunity costs, such as the time spent on the property, and potential regulations, such as taxation and registration, are insignificantly associated with the intention to become a host. Mahadevan (2019) offered several insights into the influencing factors that attract existing guests to become hosts. However, in collecting guests' viewpoints, the study could not comprehensively identify the motivations and constraints of existing hosts to continue operating in the sharing business.
Studies that systematically analyse how hosts' motivators and constraints affect hosts' continuance intention are non-existent. The present study holistically evaluates hosts' motivation factors ('motivators' in the hypotheses) and constraint factors ('constraints' in the hypotheses).
To validate the predictive power of the identified motivators and constraints, this study follows previous studies (Amaro et al., 2019;Huang and Hsu, 2009;So et al., 2018) by investigating their predictive relationships with hosts' attitudes towards P2P accommodation business ('attitude' in the hypotheses) and behavioural intention to continue their P2P accommodation business ('intention' in the hypotheses). The theory of planned behaviour (TBP) can be used to examine behavioral intention, which can predict someone's actual behavior (Ajzen 1991). When TPB is applied to evaluating tourism and hospitality cases, it is found that motivators can largely explain variation of attitude and then behavioral intention (Huang and Hsu 2009). TPB has been applied to examining determinants of guests' intention to use Airbnb (So, Oh, and Min 2018).
The negative (positive) effect of constraints (motivators) on behavioural intention will require the presence of an attitude as the explanatory variable (i.e., mediator). Specifically, constraints will lead to negative attitude towards continuance of P2P accommodation business and then intention to exit the market, whereas motivators will lead to positive attitude towards continuance of P2P accommodation business and then intention to stay in the market. Thus, the following hypotheses have been formulated.

H1:
Constraints are negatively associated with attitude.
H2: Attitude mediates the negative relationship between the constraints and intention.
H3: Motivators are positively associated with attitude.
H4: Attitude mediates the positive relationship between the motivators and intention.

Business scale as a moderator
Although research about P2P accommodation has been well-documented in the literature, studies concerning hosts' attitudes towards and intentions to operate such businesses tend to focus on predicting factors (Mahadevan, 2019;Malazizi et al., 2018;Xie et al., 2017). Less is known about the factors that moderate the predicting effects. Understanding the significant moderator can help practitioners effectively act on research implications. To fill the gap in the literature, we propose that business scale, in the form of single versus multiple properties, is a moderator.
This proposition is grounded in the size-survival relationship of organisations, which suggests that a large organisation has advantages in business continuance (Bercovitz and Mitchell, 2007). Compared with small competitors, a large business generally has advantages in terms of accessing and allocating resources (Kaul, 2012), maximising operation efficiency (Halkos and Tzeremes, 2007), bargaining power when dealing with external parties (e.g., suppliers; Tyagi, 2001), and innovativeness (Zaridis et al., 2020). Given its advantages, a large business should be more capable of addressing the constraints imposed by an environment.
However, motivators are internalised in people (Pritchard et al., 2009) and affect attitudes, which are also internalised; thus, they should be less susceptible to environmental factors (e.g., business scale). Consistent with this argument, prior study reported that entrepreneurs' motivation is primarily driven by the psychological benefits gained from operating the business (for examples independence, freedom, job satisfaction, and others) (Alstete, 2008). In sum, the motivator effect on attitude towards business continuance is unlikely to vary with business scale. By contrast, the attitude of a host is more likely to be affected by the constraints if the host operates a single property (i.e., small scale) than if the host operates multiple properties (i.e., large scale). It is hypothesised that: H5: Business scale moderates the negative relationship between the constraints and attitude, but it does not moderate the positive relationship between the motivators and attitude.
H5a: The negative relationship between constraints and attitude is significant for small-scale hosts (with a single property) but not for large-scale hosts (with multiple properties).
H5b: The positive relationship between motivators and attitude is not contingent on business scale. Figure 1 illustrates the hypotheses in a conceptual framework.

Research design
We adopted a mixed-methods approach to conduct two studies, Study 1 and Study 2 (Creswell et al., 2003). In Study 1, a qualitative phase was followed by a quantitative phase. The qualitative phase was designed to supplement the inconsistent and Western-centric results of previous research and to identify hosts' motivations and constraints, which could then be tested to provide a relevant and useful scale. The quantitative phase involved the development and analysis of a conceptual framework to demonstrate how constraints and motivations influence P2P accommodation hosts' behavioural intentions. Study 2 included further explanatory qualitative research. The unique nature of Chinese P2P accommodation hosts necessitated further exploration of the model built in Study 1. The qualitative data were additionally investigated to provide new insights that were drawn to support and explain some key findings of the quantitative research.

Study 1: Methods and findings
We used both qualitative and quantitative methods to develop a scale for Chinese hosts' motivators and constraints, following Churchill's (1979) scale development process. The items were generated based on qualitative data collected from in-depth interviews with hosts and a review of the literature (e.g., Liu and Mattila, 2017;So et al., 2018). We interviewed 20 hosts to provide rich information between June and July 2019. Themes were derived from the data (an inductive approach) based on the meaning captured in the interviews and existing literature (a deductive approach). The process identified 35 constraint items and 22 motivator items (see Table 1). The reliability and validity of these items were assessed using the questionnaire survey approach. In particular, the measurement model was tested. As discussed in the section entitled 'Motivations and constraints of P2P accommodation hosts', the predicting power of motivators and constraints was validated through an examination of their relationships with attitude and intention and the testing of the proposed hypotheses. The survey design and results are explained below.

Survey respondents and procedure
All respondents were P2P accommodation hosts who operated their businesses in China. They were asked to complete an online survey, which was distributed through 27 host social media groups on WeChat. The hosts were also asked to send the survey link to their peers. We also sent the link to our known hosts. Data collection was conducted in mid-2019 and lasted a month. By the end of the survey, we had received 334 responses. Although data were collected before the current pandemic, hosts perceived the motivations and constraints scale and the theoretical framework developed in this study to be highly relevant in providing implications for helping P2P accommodation recover following the pandemic.

Measures
This study identified 35 constraint items and 23 motivator items. They were measured on a 7point Likert scale (1 = strongly disagree, 7 = strongly agree). To measure attitudes towards the P2P accommodation business, we adapted three items from So et al. (2018). The intention to continue P2P accommodation business was operationalised in three items that had been adapted from Liu and Mattila (2017). Both attitude and intention were measured on a 7-point bipolar scale. The final measurement items are reported in Table 1. Regarding the moderator of business scale, respondents were asked if they operated a single property or multiple properties. We asked about demographic characteristics, including gender, age, education level and annual family income. Six invited experts reviewed the questionnaire prior to the pre-test. The hosts were Chinese people who were likely to have difficulty in reading English; therefore, the questionnaire was translated into Chinese. Back-translation was then conducted to ensure that the two language versions were semantically equivalent. A pre-test was conducted among four hosts who did not indicate having any problems with the questionnaire.

Measurement model
To generate the factors or dimensions for the model, we conducted a principal component analysis (PCA) of the 35 constraint and 23 motivator items separately. The analyses were performed using the direct oblimin (delta = 0) rotation method and an eigenvalue of 1 as the cutoff point. Seven constraint factors were generated (Kaiser-Meyer-Olkin = 0.889; Bartlett's test of sphericity χ 2 (595) = 6,075.27, p = 0.000; total variance explained 63.95%). Among the motivators, one item (i.e., believed the sharing economy would continue to grow) had a communality of less than 0.5 and was loaded to a factor that was not relevant (i.e., optimising the use of resources). Therefore, it was eliminated, and PCA was performed again on the remaining test of sphericity χ 2 (231) = 4,146.83, p = 0.000; total variance explained 73.05%). In the final models, all absolute values of factor loadings were greater than 0.4, and communalities were above 0.5. The factor labels are shown in Table 1. This was an exploratory study; thus, the nature of the measurement model of constraints and motivators was uncertain (i.e., reflective or formative). We empirically examined it using confirmatory tetrad analysis (CTA) in the PLS-SEM. As Becker et al. (2012) suggested, a twostage approach was adopted. First, the latent scores of the dimensions of the constraints and motivators were generated. CTA-PLS requires two criteria, which are as follows: (1) at least some indicators (i.e., dimensions) are significantly correlated and (2) more than three indicators per measurement model. The constraint dimensions were all significantly correlated with each other (p < 0.05), while 14 out of 16 correlations among the motivator dimensions were statistically significant. Therefore, the first criterion was met. Both the categories of constraints and motivators had more than three dimensions, and thus, the second criterion was met. Both constraints and motivators had at least one tetrad with a bias-corrected and Bonferroni-adjusted confidence interval not including zero, indicating that their measurement models should be formative. As such, reflective-formative HCMs of the constraints and motivators were formed and tested (see Figure 2). The use of bootstrapping estimation in the PLS-SEM ruled out the concern of data normality. First, the adequacy of the measurement model was assessed. We evaluated the internal consistency and validity of the measures based on multiple criteria. Internal consistency was demonstrated, as all composite reliability values were greater than 0.7 (see Table 2). The outer loadings were relevant, as they were above 0.4. The elimination of items with outer loadings below 0.7 did not lead to an increase in composite reliability, and, thus, they were retained. Convergent validity existed because the average variance extracted (AVE) values were greater than 0.5. Discriminant validity also existed because of three results: (1) the outer loadings of the items were greater than their cross-loadings (see Table 1); (2) the AVE values exceeded square correlations with other constructs (see Table 2); (3) the Heterotrait-Monotrait Ratio (HTMT) values were less than 0.9 (the largest value was 0.793), while the HTMTinference criterion of the bias-corrected confidence intervals did not include 1.

Common method bias
To examine if the common method bias existed, we conducted Harman's Single-Factor Test. A PCA of all items (without rotation) resulted in more than one factor, and the first factor explained 20.17% of the variance (far below 50%). The results indicated that this bias did not exist. To further validate the results, we adopted the unmeasured latent market construct method.
The conclusion remained, as (1) a few method factor loadings were statistically significant, (2) the substantive variances of items exceeded their method variances, and (3) the ratio of average substantive variance to average method variance (233:1) was large.

Structural model
The adequacy of our structural model was reflected in the non-existence of multicollinearity issues (all variance inflation factors were less than 5) and satisfactory predictive accuracy (the blindfolding test reported Q 2 values above zero).
The results of the hypothesis tests about the direct and indirect effects are shown in

The moderating effect of business scale
Business scale was a dichotomous variable (single versus multiple properties). Thus, Hypotheses 5a and 5b were examined using multi-group analysis, which requires the sample size of the bigger group to be less than twice the size of the smaller group (Hair et al. 2018). This criterion was met as there were 178 cases recorded of single-property businesses, which was less than twice the 121 cases of multiple-property businesses (i.e., 242). The analysis also requires the establishment of configural invariance and partial or full measurement invariance.
Configural invariance was demonstrated, as indicators, data coding and algorithm settings were the same for both groups. Measurement invariance was examined using a measurement Hence, partial measurement invariance was confirmed, and multi-group analysis was deemed adequate (Hair et al., 2018).
The results of the moderating effects of business scale on the constraint-attitude and motivator-attitude relationships are reported in Table 3. The permutation test, PLS-MGA, parametric test and Welch-Satterthwaite t-test converged on the same conclusion that the moderating effect exists in the constraint-attitude link (all p-values < 0.05) but not in the motivator-attitude link (all p-values > 0.05). By separately conducting structural modelling analyses for single-property and multiple-property businesses, we found that the constraintattitude link was significantly negative only for the single-property group (coefficient = -0.241, p < 0.05) and not for the multiple-property group (coefficient = 0.036, p > 0.05). The motivatorattitude link was positively significant for both groups (p < 0.05; see Table 3). Thus, H5 was supported.

Study 2: Methods and findings
The goal of the supplementary qualitative study (Study 2) was to provide interpretive evidence of the relationships among constraint, motivation, attitude, business scale and intention to continue one's P2P accommodation business in China. After the survey, we interviewed 20 hosts from different backgrounds to gather more detailed information. The duration of the interviews ranged from 25 to 45 minutes, with an average of approximately 30 minutes. The interview was audio recorded with the consent of the participants. The interviews ended when no new substantive information emerged. Two of the authors conducted the interviews and took notes and then adopted a thematic analysis to organise and understand the data (Boyatzis, 1998). In line with the survey results concerning the moderating role of business scale, the findings of Study 2 show that the owners of multiple properties were less affected by constraints.
Joyce (42, business owner, multiple properties), who owned many high-end properties in Shanghai, expressed that 'our well-designed villa is less likely to have many constraints, [un]like those who use their own home to run P2P accommodation. Our properties will bring the right clients, who want to experience natural beauty. Sometimes, we receive clothing designers' requests on the platforms. They use our villa to take photos. It is a really exciting new opportunity for us'.
All of the hosts listed their properties on various platforms, but the majority expressed that they had had a good experience with the local Chinese platforms and a relatively negative experience with Airbnb. Tom (50, engineer, single property) explained, 'It is extremely hard to contact customer services on Airbnb. They want a commission and to regulate us without delivering much. Tujia is great. Their manager teaches me how to do things and even replies to my customers. You can also transfer these features to Xiaozhu easily, but not to Airbnb.' Many studies focusing on Chinese culture have shown that Chinese people emphasise interpersonal contact (e.g., Zhang et al., 2019). The extra interpersonal services provided by local platforms have made Chinese hosts more likely to choose them. Criticisms of Airbnb have increased over its failure to adapt to Chinese culture and its limited selection of Chinese hosts compared with local platforms (Beijing News, 2020). However, some younger Chinese hosts believe that 'Airbnb is an internationally recognised brand. They bring better quality people than the local platforms do. They are fun and innovative to live with' (Lucy, 25, single property). Nevertheless, the perceived special interpersonal relations might be a unique selling point, differentiating the local Chinese platforms from Airbnb.

Conclusions
The sharing economy has become an important and popular research topic (Heo, 2016). The findings of the present study contribute to the further development of a theoretical understanding of the sharing economy. Using a mixed-methods approach with two studies, this paper captures a comprehensive set of motivation and constraint factors facing hosts and builds a theoretical framework to validate the predictive power of these factors on hosts' attitudinal and behavioural responses. Qualitative data provide several in-depth explanations for the key findings generated from the questionnaire survey. The paper contributes to the formation of a holistic framework for analysing the motivations and constraints affecting continuity in a key sector of the sharing economy, P2P accommodation. Although the factors that encourage or deter guests from staying in P2P accommodation have been well examined in the literature (e.g., So et al., 2018;Tussyadiah, 2015),

Theoretical implications
This research enriches the discussion on the sharing economy through (a) testing factors captured in previous studies and (b) uncovering additional key drivers and deterrents. The findings of our semi-structured in-depth interviews revealed various themes of constraint and motivation. The additional qualitative findings provided interpretive evidence to support and explain the survey results. The hosts' perceived constraints included themes that have been identified in previous research, such as guest issues, government regulations and market competition (e.g., Alraeeini et al., 2019;Malazizi et al., 2018). The findings on P2P platforms constraints, including the difficulty of rejecting a booking due to platform regulations, and inconsistency in the policy/requirements of different platforms, align with those of Farmaki (2019), who noticed that Airbnb's inconsistent regulations and favouring of guests presented challenges to hosts seeking to reject bookings. Our results uncovered new items, such as the need to put much effort into maintenance and cleaning.
Among the motivators, better interpersonal relations, settling the cost of living, optimising the use of resources and ease of operation have been reported in the literature (e.g., Farmaki and Stergiou, 2019;Karlsson and Dolnicar, 2016). The result of offering more autonomy, freedom and flexibility is consistent with Benoit et al. (2017). Jung et al. (2016) found that having a 'human relationship' becomes a key shared asset of P2P accommodation, but making a profit remains important. Our findings uncovered motivators that have not been discussed in the literature, including showcasing one's trendy character, such as feeling it was 'a cool thing' or 'a novel thing' to run a sharing-economy business, and being one's own boss.
Most existing studies have investigated P2P accommodation hosts operating on the Airbnb platform in the West, such as in Northern Cyprus (Malazizi et al., 2018), the United States (Ert and Fleischer, 2019) and Europe (Farmaki and Stergiou, 2019). This paper is among the few that explore hosts operating P2P accommodation on global and local platforms in Asian countries.
Thus, it generates several new ideas about motivations and constraints; the perception that it is 'a cool/novel thing' to run a sharing-economy business appears to be one of the most important motivators in this context. Previous studies have reported that sharing and interaction are the key drivers of joy (Karlsson and Dolnicar, 2016). Complementing this aspect of the literature, our study found that considering the operation of P2P accommodation to be 'a cool/novel thing' made Chinese hosts feel good. Our findings also reveal that hosts face new P2P platform constraints, including a lack of technical support and non-responsiveness to hosts' inquiries.
These issues may be due to a lack of professional workers on platforms or hosts' unfamiliarity with the platforms and relevant regulations. Chinese hosts also felt that having to put too much effort into maintenance was a severe constraint. This finding differs from Mahadevan (2019), who found that opportunity costs, including time and effort investment in maintenance, was not significantly associated with the intention to become a host based on a survey among Australians. This might be because, unlike Western hosts, most Chinese are not used to engaging in accommodation maintenance by themselves.
The paper enriches the literature by developing a scale to measure the motivators and constraints of operating P2P accommodation. The predictive power and validity of motivators and constraints were demonstrated in the relationships with attitude and behavioural intention.
The advantage of this empirical examination is that it took the influence of both motivation and constraint factors into account on determining outcome variables. The results indicate that constraints lower the intention to continue one's business because of an unfavourable attitude, whereas motivators heighten the intention because of a favourable attitude.
The simultaneous examination of motivators and constraints enabled the comparison of their effects on attitude and intention. We found that in general, motivators had a greater effect on attitude and intention than did constraints. Motivations play a more important role than do constraints in influencing hosts' attitudes towards P2P sharing accommodation. The findings are not surprising if the moderating effect of business scale is considered. The positive effect of motivators on attitude is positive and robust against business scale, whereas the negative effect of constraints on attitude only exists for single-property hosts. In line with the size-survival relationship (Bercovitz and Mitchell, 2007), we found that constraints have a greater impact on small businesses (i.e., single properties) owing to their insufficient resources to address the problems arising from the constraints. Therefore, single-property hosts hold a greater negative attitude towards operating a P2P accommodation. The motivator effect is not contingent on business scale because motivation is internalised to the host. This internalised trait has low relevancy with the external factor of business scale. Their interaction did not influence attitude.
The findings of the moderating effect add knowledge to the literature by showing the asymmetrical effects of constraints and motivators on their consequences. Future research should examine the boundary situation qualifying the constraint and motivator effects.

Practical implications
This study identifies several constraints on operating P2P accommodation. Working with the relevant industry associations, platforms could provide self-learning packages and training to enable hosts to familiarise themselves with the legislative status of the P2P accommodation business and improve their communication skills with guests. Platforms could also work with third parties to offer services, such as maintenance and cleaning, to reduce hosts' burden. Finally, there should be measures to reinforce the motivations recognised by hosts. For example, platforms could provide opportunities for offline interactions between guests and hosts, especially for those who value interpersonal relations as a motivator. Brands that showcase the trendy characterisations of the sharing accommodation business could be designed to attract new hosts.
The results have two types of policy implication for government departments. First, government regulations, such as the intervention and accreditation requirements of the local authorities and the charging of taxes, pose significant constraints for existing hosts. P2P accommodation hosts are suffering from an increase in expenses and dramatic decrease in revenues due to COVID-19 (Dolnicar and Zare, 2020). Unclear government regulations seem to be an underlying constraint on sharing accommodation, especially in the context of this crisis.
Therefore, it is crucial to design new regulatory measures (Dolnicar, 2019) to standardise rules for this novel market and offer support for its sustainable development. Second, the other major concerns for hosts are weak market demand and stiff market competition. Similar to traditional tourism and hospitality businesses, the sharing accommodation industry is sensitive to the external market environment. Tourism bureaus and hospitality associations should incorporate this business into their marketing strategies and promote it with unique and positive images. considered.
This paper also has novel insights for existing and future hosts. They should measure prospective gains against costs according to the holistic framework developed in this paper before deciding to join, continue or exit the P2P sharing business. Hosts need to be fully aware of the constraints and alert to potential problems, such as regulations, guest issues and seasonality, so that they can prepare the corresponding solutions. Hosts should carefully consider the recommendations for addressing constraints, especially if they operate a single property, given the moderating effect of business scale.

Limitations and future research
The implications of this study should be considered in light of several limitations. First, despite our efforts to recruit diverse participants, the population of Chinese hosts was unknown. We might have missed hosts who were resistant to taking part in the study. Future studies could try to extend the sample to reach hosts in other Asian countries, where local platforms are also popular. Second, the P2P platform constraint reflected respondents' generic experiences with various online platforms. The variances aroused from evaluations based on global or local platforms were unknown. Future research is recommended to measure this constraint factor by asking the respondents to base their evaluation on a particular platform or the one that they most recently used. Third, the COVID-19 pandemic is likely to have altered hosts' perceptions on motivations and constraints that were not captured in this research, which was conducted before the pandemic. However, the measurement scale and model developed in this paper can provide a useful framework for future research to explore the motivators and constraints caused by the pandemic. This research examined hosts operating sharing accommodations in China. A review of the literature showed that most studies have focused on hosts in Western countries; thus, the findings of this paper are both general and unique constraints and motivations. Future research could examine the similarities and differences in the attitudes and behavioural intentions of hosts in Western and Eastern countries.