1 INTRODUCTION

Tourism ranks among the most important economic activities in the world. Recent figures from the World Tourism Organisation (UNWTO, 2018) show that the number of international tourist trips reached nearly 1.4 billion in 2018, generating an economic impact estimated in US$ 1.7 billion (and 10 per cent of the jobs in the world). Despite the fast growth of many competitive destinations, Europe is still the main tourist destination with 51% of arrivals and 39% of the total income generated by international tourism.

However, there is also consensus about the substantial negative impacts of international tourism both at a global scale (mainly air transport greenhouse gas emissions) and on local communities (i.e., environmental externalities such as pollution, congestion and resource depletion). As tourism depends heavily on the physical environment but also has a significant impact on it (Tisdel, 2005), tourism sustainability has become a central issue in the last decades. A substantial academic debate has focused on the use of measures and instruments to protect the environment. Under some circumstances (Pazienza, 2011; Piga, 2003; Rinaldi, 2014; Schubert, 2010), the use of tourist related taxes might constitute an efficient way to cope with tourism related externalities. Partly for these reasons, the use of environmental taxation in this sector has been receiving increasing attention in the last decades (Backhaus, 1999; Barde & Owens, 1996; Cirer Costa & De Ibiza, 2008; Corthay & Loeprick, 2010; Gago, Labandeira, Picos, & Rodríguez, 2006; Gago, Labandeira, Picos, & Rodríguez, 2009; Gooroochurn, 2004; Hughes, 1981; Rotaris & Carrozzo, 2019; Sheng, 2011; Sheng & Tsui, 2009; Tavares, 2011).

A great variety of taxes is applied in the tourism industry (Durán Román, Cárdenas García, & Pulido Fernández, 2020; García, Marchena, & Morilla, 2018; Gooroochurn & Sinclair, 2005) and these can be applied by governments to raise funds to offset the environmental impacts of tourism, as shown by Durán Román et al. (2020) in their review of tourism taxes in the top 50 tourist destinations. The OECD (2014, page 76) defines tourism taxes as ‘the indirect taxes, taxes and tributes that mainly affect the activities related to tourism’ and classifies them into the following types:

1. Arrival and departure taxes (taxes and fees that apply to the departure or arrival of a country, such as visas and fees for transit and movement of passengers or crew);
2. Air travel taxes payable at airports and normally applied to airline trips to cover costs arising from the service provided at airports;
3. Hotels and accommodation taxes (with different names such as overnight-tax, lodging-tax, bed-tax, occupancy-tax, room-tax, or accommodation-tax) payable at the accommodation and usually managed by regional or local administrations;
4. Taxation on consumption, such as value added tax (VAT) or other taxes on goods and services (although taxation affects tourists and residents, differentiated tax treatment is sometimes applied to tourism goods and services);
5. Environmental taxes: although the use of environmental taxes or eco-taxes is generalised, sometimes there are specific environmental taxes for tourist activities that attempt to protect natural or cultural spaces of special tourist value; and
6. incentives: tax reductions and exemptions for certain activities that the government wishes to promote.

Accommodation taxes (i.e., taxes paid directly by the client at the hotel, charged for each night of accommodation), are widely applied specially in the main cities of Europe and are considered the most common example of tourist-related taxes. Notwithstanding, general indirect taxes may be preferred to specific taxes as a more feasible, equitable and neutral way of obtaining tax revenues from tourism activities (Gago et al., 2009). Nevertheless, when taxes are justified to overcome the negative effects of tourism activities, Roselló and Sansó (2017) make the point that the term tourist tax does not refer to a tax solely paid by tourists, but mainly paid by them (Dwyer, Forsyth, & Dwyer, 2010). In this way, increases in general taxes (such as VAT) are criticised, as many of the tourism-related goods are consumed also by residents who would be paying for external costs that they do not create.

The distribution of the tax burden and the success of a tourism tax policy is also affected by the characteristics of the market. For example, Barnett (1980) shows that the benefit of an optimal tax on a monopoly can be less than the damage it causes. In fact, a higher tax would induce the monopolist to further reduce the level of production and consequently the consumer surplus. Sheng (2017), on the other hand, advises lower taxes as the market becomes more competitive and demand more price-elasticity (due to the difficulty of passing on the tax to tourists), although the existence of negative externalities may make an increase in taxes advisable. Under a classical Pigouvian approach, an optimal tax should be equal to the external marginal damage at the point where social marginal benefits equal social marginal costs (Pearce and Turner, 1989). Nevertheless, difficulties in the estimation of benefits and costs makes it difficult to apply optimal Pigouvian taxes and sub-optimal solutions have been suggested (Baumol and Oates, 1971), consisting of using cost-efficient taxes to achieve certain environmental quality standards (not necessarily optimal in the Pigouvian sense).

In the case of tourism, the existence of negative externalities is the main argument for the use of tourism taxes (OECD, 1989) as a tool to internalise the costs derived from an excessive boom in visitors. In the case of large cities, with unique attractions and a high degree of visitor loyalty (such as Venice or Barcelona), studies show limited effectiveness from tourist accommodation taxes for this purpose, due to the low elasticity of demand (Heffer-Flaata, Voltes-Dorta, & Suau-Sanchez, 2020). However, this price sensitivity increases in those destinations with less differentiation and better substitutes (typically sun and beach destinations), where demand is more sensitive to price. Thus, the response of tourists to an accommodation tax is specific to each place, which would make it advisable to carry out specific studies of preferences for each destination.

Another important reason used to justify the use of accommodation taxes, in addition to reducing the inefficiencies generated by negative externalities, is that they have proven to be an important source of revenue for governments that can be raised from non-residents and used to improve the attractiveness of the destination (Gooroochurn & Sinclair, 2005; Palmer & Riera, 2003; Sheng & Tsui, 2009).

On the contrary, the main argument of the tourism industry against this type of tax is centred on the possible loss of competitiveness. As Collins and Stephenson (2018) point out, tourists will seek alternative destinations at a lower price, which can have a significantly negative effect in those destinations highly dependent on tourism. For example, Lee (2014) argued that price increases due to a bed tax for Midlands Hotels, in the competition between geographic submarkets, would imply a significant loss of competitiveness. Also, even if an increase in the price of accommodation does not significantly reduce demand for accommodation, it may reduce demand for other goods given the composite nature of tourism demand.

Additionally, as Aguiló, Riera, and Rosselló (2005) point out, the contraction in tourism demand may be greater if the destination of the tax revenues does not respond to the preferences of tourists or if these preferences are not used to improve the attractiveness of the destination (Rotaris and Carozzo, 2019). The level of acceptance of the tax is affected by factors such as the perceived fairness of the amount collected (Chung, Kyle, Petrick, & Absher, 2011) as well as the existence of a clear explanation of the collection mechanism and how the amount of tax collected will be used (Rotaris and Carozzo, 2019). Some studies show that tourists are more willing to pay taxes when they are linked to environmental outcomes (Cantallops, 2004; Lee & Pearce, 2002), rather than general taxes for tourism development (Edwards, 2009). Other studies conclude that willingness to pay a tax for tourist accommodation depends on the profile of the tourist. Thus, according to Do Valle, Pintassilgo, Matias, and André (2012) in the case of the Algarve (Portugal), tourists who showed a lower willingness to pay corresponded to the segment of ‘typical sun and beach tourists’, while the willingness to pay increased as the tourist profile became more environmentally friendly. Thus, the analysis of tourists' preferences is one of the most important elements in the success of a tourism rate. Another important issue is the analysis of the different preferences between residents and tourists, a comparison that is not generally addressed in the literature, with some exceptions such as the studies by Brau, Scorcu, and Vici (2009) and Figini, Castellani, and Vici (2009) for the case of Rimini in Italy.

In relation to the methodology used to analyse tourism taxes (see Biagi, Brandano, & Pulina, 2017) we can find different types of models. A first type uses panel data or time series (Aguiló et al., 2005; Arguea & Hawkins, 2015; Heffer-Flaata et al., 2020) which allow for the estimation of tourism demand elasticities and the prediction of the effects on tourism flows of the imposition of a tax; however, they have the disadvantage of requiring a series of data over a long period of time. Other types of models analyse the impacts of taxes from a macroeconomic perspective considering intersectoral relations using input–output or general equilibrium models (Forsyth, Dwyer, Spurr, & Pham, 2014; Gago et al., 2009; Ihalanayake, 2012; Ponjan & Thirawat, 2016; Sheng & Tsui, 2009). Although these are static analyses, these types of models also require large datasets and are not applicable for analysing the effects of applying a tourism rate in a specific location.

Most studies analysing the effects of tourism taxes have been carried out after they have been implemented, so they are ex-post analyses the conclusions of which cannot be applied to the design of the tax. The optimal design of a tax requires estimating preferences and the response of tourists before applying it, which advises the use of stated preference techniques such as discrete choice models. Since scenarios in discrete choice models can be constructed with hypothetical attributes, this technique allows for the estimation of the reaction to certain policies before they are implemented. However, the use of these techniques is rare, with a few exceptions such as the study by Rotaris and Carrozzo who applied contingent valuation to estimate the willingness to pay for tourism taxes according to the use of tax revenues in the Italian context.

In the case of tourism, discrete choice models have been applied in different areas (see the review by Crouch & Louviere, 2000); for example, to model individual preferences on attributes of the demand for tourism goods or services (Boto-García, Mariel, Pino, & Alvarez, 2020; Chen, Masiero, & Hsu, 2019; Masiero, Heo, & Pan, 2015) the choice of destinations (Huybers, 2003), the environmental and recreational value of cultural or natural heritage (Lee, Mjelde, Kim, Lee, & Choi, 2019; Lupu, Padhi, Pati, & Stoleriu, 2020; Perez Loyola, Wang, & Kang, 2019), or the effects of certain policies, such as environmental conservation policies (see for example, Bocci, Sohngen, Lupi, & Milian, 2020; Xu et al., 2020). Nevertheless, Rosselló and Sansó (2017) noted two difficulties in applying choice models to assess the effect of a tax increase on the international demand for a tourist destination: (a) the use of the distance variable as an indicator of price, makes it impossible to analyse price increases as a result of the tax, and (b) the difficulty of having a representative sample of international tourists with information about alternative destinations. These difficulties may explain the scarcity of studies using discrete choice models to analyse the effects of tourism taxes before applying them to a specific location.

This review of literature allows us to conclude that the design of a new accommodation tax would be greatly enhanced by having a previous study about the preferences of both tourists and residents, specific to the destination. Discrete choice models also allow us to obtain valuations of the different attributes of a tax and of the preferred destination of the funds collected, which could condition the degree of acceptance of the tax. To the best of our knowledge, there is a scarcity of studies with these characteristics, so the analysis of a possible room tax opens up the possibility of making a valuable contribution to the existing literature on this subject.

This paper analyses the public's response to a hypothetical room-tax applied to overnight stays in the Canary Islands. For this we used an opinion survey, based on a Likert-scale questionnaire, and a discrete choice experiment to estimate the willingness to pay for a room-tax and the factors that determine its acceptability.

The methodology was applied separately to tourists and residents of the islands, allowing for a comparison of their preferences. Specifically, we wanted to answer the following questions:

1. under what conditions would tourists/residents be willing to accept the introduction of a new tax and, thus,
2. how should the tax be designed to better meet both the preferences of tourists and residents?

The remainder of the paper is structured as follows: Section 2 reviews the use of room taxes in Spain; the methodology and fieldwork are explained in Section 3; Section 4 summarises the main findings from the opinion survey; Section 5 presents the main results of our stated choice survey and accompanying discrete choice model estimation and, finally, Section 6 reveals our main conclusions.

中文翻译：

---

加那利群岛有房税的余地吗？

1 介绍

旅游业是世界上最重要的经济活动之一。世界旅游组织（UNWTO，2018 年）的最新数据显示，2018 年国际旅游人数达到近 14 亿人次，产生的经济影响估计为 17 亿美元（占世界就业机会的 10%）。尽管许多有竞争力的目的地增长迅速，但欧洲仍然是主要的旅游目的地，占国际旅游总收入的 51% 和 39%。

然而，对于国际旅游在全球范围内（主要是航空运输温室气体排放）和对当地社区（即环境外部性，如污染、拥堵和资源枯竭）的重大负面影响，也存在共识。由于旅游业严重依赖自然环境，但也对其产生重大影响（Tisdel，2005 年），因此旅游业的可持续性已成为过去几十年的核心问题。一场实质性的学术辩论集中在保护环境的措施和工具的使用上。在某些情况下 (Pazienza, 2011 ; Piga, 2003 ; Rinaldi, 2014 ; Schubert, 2010)，旅游相关税收的使用可能是应对旅游相关外部性的有效方式。部分由于这些原因，在过去几十年中，该部门对环境税的使用受到越来越多的关注（Backhaus，1999 年；Barde & Owens，1996 年；Cirer Costa & De Ibiza，2008 年；Corthay & Loeprick，2010 年；Gago，Labandeira ，皮科斯德，和罗德里格斯，2006年;加戈，Labandeira，皮科斯德，和罗德里格斯，2009年; Gooroochurn，2004年，休斯，1981 ; Rotaris＆Carrozzo，2019 ;盛，2011 ;盛和翠，2009年，塔瓦雷斯，2011）。

旅游业征收的税款种类繁多（Durán Román、Cárdenas García 和 Pulido Fernández，2020 年；García、Marcena 和 Morilla，2018 年；Gooroochurn 和 Sinclair，2005 年），政府可以应用这些税收来筹集资金抵消旅游业对环境的影响，如 Durán Román 等人所示。( 2020 ) 在他们对前 50 名旅游目的地的旅游税的审查中。经合组织（2014 年，第 76 页）将旅游税定义为“主要影响与旅游相关活动的间接税、税收和贡品”，并将其分为以下几类：

1. 到达和离开税（适用于出发或到达一个国家的税费，例如签证以及乘客或机组人员的过境和流动费用）；
2. 在机场应付的航空旅行税，通常适用于航空旅行，以支付机场提供的服务产生的费用；
3. 在住宿处缴纳的酒店和住宿税（有不同名称，如过夜税、住宿税、床位税、占用税、房间税或住宿税），通常由地区或当地政府管理；
4. 消费税，如增值税（VAT）或其他商品和服务税（虽然税收影响游客和居民，但有时对旅游商品和服务实行差异化税收待遇）；
5. 环境税：虽然环境税或生态税的使用是笼统的，但有时会对试图保护具有特殊旅游价值的自然或文化空间的旅游活动征收特定的环境税；和
6. 激励措施：政府希望促进的某些活动的税收减免。

住宿税（即由客户直接在酒店支付，按每晚住宿收取的税费）在欧洲主要城市被广泛应用，被认为是与旅游相关的税费最常见的例子。尽管如此，作为从旅游活动中获得税收的一种更可行、公平和中立的方式，一般间接税可能比从量税更受青睐（Gago 等，2009 年）。然而，当有理由通过税收来克服旅游活动的负面影响时，Roselló 和 Sansó（2017 年）指出，旅游税一词并不是指仅由游客支付的税款，而是主要由游客支付的税款（Dwyer、Forsyth，和德怀尔，2010）。通过这种方式，一般税收（如增值税）的增加受到批评，因为许多与旅游相关的商品也被居民消费，他们将支付他们没有创造的外部成本。

税收负担的​​分配和旅游税收政策的成败也受市场特点的影响。例如，Barnett ( 1980 ) 表明最优税收对垄断企业的收益可能小于其造成的损害。事实上，更高的税收会促使垄断者进一步降低生产水平，从而降低消费者剩余。盛 ( 2017)，另一方面，随着市场变得更加竞争和需要更多的价格弹性（由于难以将税收转嫁给游客），建议降低税收，尽管负外部性的存在可能使增加税收是可取的。在经典庇古方法下，当社会边际收益等于社会边际成本时，最优税收应等于外部边际损害（Pearce and Turner，1989）。然而，收益和成本估算的困难使得应用最优庇古税变得困难，并且已经提出了次优解决方案（Baumol 和 Oates，1971 年），包括使用具有成本效益的税收来实现某些环境质量标准（不一定庇古意义上的最优）。

在旅游业的情况下，负外部性的存在是使用旅游税（经合组织，1989 年）作为工具来内部化因游客过度繁荣而产生的成本的主要论据。对于具有独特景点和高度游客忠诚度的大城市（例如威尼斯或巴塞罗那），研究表明，由于需求弹性低，因此为此目的征收旅游住宿税的效果有限（Heffer-Flaata、Voltes -多尔塔和苏奥-桑切斯，2020）。然而，这种价格敏感性在那些差异较小且替代品更好的目的地（通常是阳光和海滩目的地）中增加，在这些目的地，需求对价格更敏感。因此，游客对住宿税的反应因地而异，因此最好对每个目的地的偏好进行具体研究。

除了减少负外部性产生的低效率外，另一个用来证明使用住宿税合理的重要原因是，它们已被证明是政府的重要收入来源，可以从非居民那里筹集并用于改善目的地的吸引力（Gooroochurn & Sinclair, 2005 ; Palmer & Riera, 2003 ; Sheng & Tsui, 2009）。

相反，旅游业反对这种税收的主要论点集中在可能丧失竞争力上。正如柯林斯和斯蒂芬森 ( 2018 ) 指出的那样，游客将以较低的价格寻找替代目的地，这会对高度依赖旅游业的目的地产生显着的负面影响。例如，Lee ( 2014 ) 认为，在地理子市场之间的竞争中，由于对 Midlands 酒店征收床位税而导致的价格上涨将意味着竞争力的显着丧失。此外，即使住宿价格上涨不会显着减少住宿需求，但鉴于旅游需求的复合性质，它可能会减少对其他商品的需求。

此外，正如 Aguiló、Riera 和 Rosselló ( 2005 ) 指出的那样，如果税收收入的目的地不响应游客的偏好，或者如果这些偏好没有被用来提高旅游吸引力，那么旅游需求的收缩可能会更大。目的地（Rotaris 和 Carozzo，2019 年）。税收的接受程度受诸如征收金额的感知公平性等因素的影响（Chung、Kyle、Petrick 和 Absher，2011 年）以及是否存在对征收机制的清晰解释以及如何征收将使用征收的税款（Rotaris 和 Carozzo，2019 年）。一些研究表明，当游客与环境结果相关时，他们更愿意纳税（Cantallops，2004 年；Lee & Pearce，2002 年），而不是旅游发展的一般税收（爱德华兹，2009 年）。其他研究得出结论，为旅游住宿缴税的意愿取决于旅游者的情况。因此，根据 Do Valle、Pintassilgo、Matias 和 André（2012 年）在阿尔加维（葡萄牙）的情况下，表现出较低支付意愿的游客对应于“典型的阳光和海滩游客”部分，而随着游客形象变得更加环保，支付意愿增加。因此，对游客偏好的分析是旅游率成功的最重要因素之一。另一个重要问题是对居民和游客之间不同偏好的分析，这种比较在文献中一般没有涉及，但有一些例外，例如 Brau、Scorcu 和 Vici（2009 年）以及 Figini、Castellani 和 Vici 的研究( 2009 ) 意大利里米尼案。

关于用于分析旅游税的方法（参见 Biagi、Brandano 和 Pulina，2017 年），我们可以找到不同类型的模型。第一类使用面板数据或时间序列（Aguiló 等人，2005 年；Arguea & Hawkins，2015 年；Heffer-Flaata 等人，2020 年），它允许估计旅游需求弹性并预测对旅游流量的影响征收税款；然而，它们的缺点是需要长时间的一系列数据。其他类型的模型使用投入产出或一般均衡模型从宏观经济角度分析税收的影响（Forsyth、Dwyer、Spurr 和 Pham，2014 年；Gago 等人，2009 年；伊哈拉纳亚克，2012 年；Ponjan 和 Thirawat，2016 年；盛和崔，2009 年）。虽然这些是静态分析，但这些类型的模型也需要大型数据集，不适用于分析在特定地点应用旅游率的影响。

大多数分析旅游税影响的研究都是在实施后进行的，因此是事后的分析其结论不能应用于税收设计。税收的优化设计需要在应用之前估计偏好和游客的反应，这建议使用陈述偏好技术，如离散选择模型。由于离散选择模型中的场景可以用假设属性构建，因此该技术允许在实施之前估计对某些政策的反应。然而，这些技术的使用很少，只有少数例外，例如 Rotaris 和 Carrozzo 的研究，他们应用条件估值来估计根据意大利背景下税收收入的使用情况来估计支付旅游税的意愿。

就旅游业而言，离散选择模型已应用于不同领域（参见 Crouch & Louviere, 2000 的评论）；例如，根据旅游商品或服务需求的属性对个人偏好进行建模（Boto-García、Mariel、Pino 和 Alvarez，2020 年；Chen、Masiero 和 Hsu，2019 年；Masiero、Heo 和 Pan，2015 年）目的地的选择（Huybers，2003 年）、文化或自然遗产的环境和娱乐价值（Lee、Mjelde、Kim、Lee 和 Choi，2019 年；Lupu、Padhi、Pati 和 Stoleriu，2020 年；Perez Loyola、Wang 和康, 2019)，或某些政策的效果，例如环境保护政策（例如，参见 Bocci、Sohngen、Lupi 和 Milian，2020 年；Xu 等人，2020 年）。尽管如此，Rosselló 和 Sansó ( 2017) 指出在应用选择模型来评估税收增加对旅游目的地国际需求的影响时有两个困难：(a) 使用距离变量作为价格指标，因此无法分析价格上涨税，以及 (b) 难以获得具有替代目的地信息的国际游客的代表性样本。这些困难可能解释了在将旅游税应用于特定地点之前，使用离散选择模型来分析旅游税影响的研究的稀缺性。

通过对文献的回顾，我们可以得出结论，通过先前对特定目的地的游客和居民的偏好进行研究，将大大加强新住宿税的设计。离散选择模型还允许我们获得税收的不同属性和所收集资金的首选目的地的估值，这可能会影响税收的接受程度。据我们所知，具有这些特征的研究很少，因此对可能的房间税的分析开辟了对有关该主题的现有文献做出宝贵贡献的可能性。

本文分析了公众对适用于加那利群岛过夜住宿的假设房税的反应。为此，我们使用了基于李克特量表问卷的意见调查和离散选择实验来估计支付房税的意愿以及决定其可接受性的因素。

该方法分别适用于岛上的游客和居民，可以比较他们的偏好。具体来说，我们想回答以下问题：

1. 在什么条件下游客/居民愿意接受新税的引入，从而，
2. 应如何设计税收以更好地满足游客和居民的偏好？

本文的其余部分结构如下：第 2 部分回顾了西班牙房间税的使用情况；方法论和实地考察在第 3 节中解释；第 4 节总结了意见调查的主要发现；第 5 节介绍了我们陈述的选择调查和随附的离散选择模型估计的主要结果，最后，第 6 节揭示了我们的主要结论。