Resilience Firm Survey 2020 - Antigua and Barbuda, Bahamas, Barbados...and 10 more

Abstract
---------------------------

The RFS in Caribbean was conducted in 13 countries between March and November 2020 and focused on the tourism industry and the restaurant, hotel and tour and transport companies. Due to the COVID-19 crisis, data collection was done both remotely and in-person depending on the restrictions in place and preference of respondent. The countries covered included Antigua and Barbuda, Bahamas, Barbados, Dominica, Dominican Republic, Grenada, Jamaica, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Sint Maarten, Trinidad and Tobago, Turks and Caicos. The survey in the Caribbean focused on impacts of recent disasters to have affected the region, including Hurricane Irma, Hurricane Maria, Tropical Storm Dorian, etc. (see Table 2 for country and disaster list). The data collection was financed by the Global Facility for Disaster Reduction and Recovery (GFDRR) with the objective of better understanding how natural hazards – large and small, affect the tourism industry in the Caribbean. The data informed the 360° Resilience: A Guide to Prepare the Caribbean for a New Generation of Shocks (Rozenberg, et al. 2021) to make recommendations on how Caribbean countries can invest resources to strengthen resilience in the region.

This project was a collaborative effort between GFDRR and Urban, Disaster Risk Management, Resilience and Land Global Practice (GPURL).

Geographic coverage
---------------------------

Antigua and Barbuda, Bahamas, Barbados, Dominica, Dominican Republic, Grenada, Jamaica, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Sint Maarten, Trinidad and Tobago, Turks and Caicos.

Analysis unit
---------------------------

- Firm level

Kind of data
---------------------------

Sample survey data [ssd]

Sampling procedure
---------------------------

The sample was drawn to achieve representativeness at the country level as well as the regional level. In the Dominican Republic, sampling was done in a way to also achieve representativeness in 4 provinces in the country. Since there was no comprehensive list of firms operating in the tourism industry readily available to sample from, the firm hired to collect data created a sampling frame from scratch by contacting relevant organizations and websites. To be able to say something about different sectors within the tourism industry, the sampling was stratified by three sectors, including hotels and accommodation, restaurants and bars, and a third sector including rental, taxi and tour companies, attractions and souvenir shops (referred to in this note as hotel, restaurant and tour/transport sectors). The sample selection was then completed in one stage in which firms were selected by using a systematic random sampling method from each stratum.

Once the firm is selected for inclusion in the survey, every effort was made to interview the firm. The survey response rate was low due to the COVID pandemic, and replacements were done. Replacements were drawn from the same stratum. Due to restrictions in some countries, firms were not reachable, even after several attempts and replacements had been done. To compensate for low response rate in some countries, the sample size in other countries was increased. As a result, The Bahamas and Turks and Caicos have lower than expected sample size so caution should be applied when interpreting country level results from these two countries. See Technical Note for more detail on composition of final sample.

The final sample contains a total of 1413 firms across the 13 countries. Dominican Republic has the largest number of observations because the objective of sampling was also to achieve province level representativeness, in addition to country level representativeness, in 4 providences that rely heavily on tourism.

To make the survey estimates representative of the population, it is necessary to apply weights to selected firms during analysis. Regional weights (weight) are applied to statistics representing regional values while country weights (weight_i) are applied to all country level statistics.

Mode of data collection
---------------------------

Computer Assisted Personal Interview [capi]

Research instrument
---------------------------

• Respondent characteristics
• Firm characteristics
• Clients
• Infrastructure dependence and disruptions
o Water
o Electricity
o Communication (phone and internet)
o Road and boat
• Suppliers
• Disaster preparedness
• Impacts of recent disasters (see Table 2)
• Impacts of disease outbreaks (Zika and COVID-19)
• Financial accounts

Cleaning operations
---------------------------

The following data editing was done for anonymization purpose:
• Precise location data, such as GPS coordinates, and subnational administrative divisions (admin 1) were dropped
• Identifying and contact information, such as firm name, respondent’s name, supplier names, phone number and email contact, were dropped
• Number of fulltime workers above 100 was recoded to “above 100 fulltime workers” to mitigate re-identification of the largest firms.
See technical note for more details on anonymization.

摘要
---------------------------
加勒比地区的风险财务调查（RFS）于2020年3月至11月间在13个国家进行，聚焦于旅游业以及餐馆、酒店以及旅游和交通公司。鉴于COVID-19疫情的影响，数据收集采取了远程和实地相结合的方式进行，具体取决于当时实施的限制措施和受访者的偏好。调查覆盖的国家包括安提瓜和巴布达、巴哈马、巴巴多斯、多米尼克、多米尼加共和国、格林纳达、牙买加、圣基茨和尼维斯、圣卢西亚、圣文森特和格林纳丁斯、圣马丁、特立尼达和多巴哥、特克斯和凯科斯群岛。加勒比地区的调查重点关注近期影响该地区的一系列灾害，包括艾尔玛飓风、玛丽亚飓风、多里安热带风暴等（详见表2中的国家和灾害列表）。数据收集由全球灾害减少与恢复设施（GFDRR）资助，旨在更好地理解自然灾害——无论大小——如何影响加勒比地区的旅游业。这些数据为《360°韧性：为新一代冲击做好准备的海地指南》（Rozenberg等，2021年）提供了信息，以提出关于加勒比国家如何投资资源以增强区域韧性的建议。

本项目是GFDRR与城市、灾害风险管理、韧性和土地全球实践（GPURL）之间的合作成果。

地理覆盖范围
---------------------------
安提瓜和巴布达、巴哈马、巴巴多斯、多米尼克、多米尼加共和国、格林纳达、牙买加、圣基茨和尼维斯、圣卢西亚、圣文森特和格林纳丁斯、圣马丁、特立尼达和多巴哥、特克斯和凯科斯群岛。

分析单元
---------------------------
- 企业层面

数据类型
---------------------------
样本调查数据 [ssd]

抽样程序
---------------------------
样本的抽取旨在实现国家层面以及区域层面的代表性。在多米尼加共和国，抽样方法旨在实现国家层面和4个省（这些省份对旅游业高度依赖）层面的代表性。由于没有现成的、涵盖旅游业运营公司的完整清单可供抽样，负责收集数据的企业通过与相关组织和网站联系，从头开始创建抽样框架。为了能够对不同旅游行业的部门发表意见，抽样根据三个部门进行分层，包括酒店和住宿、餐馆和酒吧，以及第三个部门包括租赁、出租车和旅游公司、景点和纪念品商店（在本报告中称为酒店、餐馆和旅游/交通部门）。然后，在第一阶段完成样本选择，其中每个层级通过系统随机抽样方法选择企业。

一旦企业被选入调查，便尽力进行访谈。由于COVID大流行导致调查响应率较低，因此进行了替换。替换样本从同一层级抽取。由于某些国家的限制，即使经过多次尝试和替换，企业也无法联系。为了弥补某些国家响应率低的情况，其他国家的样本量增加。因此，巴哈马和特克斯和凯科斯群岛的样本量低于预期，因此在解读这两个国家的国家层面结果时应谨慎。有关最终样本构成的更多详细信息，请参阅技术说明。

最终样本包含13个国家共1413家企业的数据。多米尼加共和国拥有最多的观测值，因为抽样目标不仅是要实现国家层面的代表性，还要在4个高度依赖旅游业的省份实现省层面的代表性。

为了使调查估计值代表总体，在分析过程中需要对选定的企业应用权重。区域权重（权重）应用于代表区域价值的统计数据，而国家权重（权重_i）应用于所有国家层面的统计数据。

数据收集方式
---------------------------
计算机辅助个人访谈 [capi]

研究工具
---------------------------
• 受访者特征
• 企业特征
• 客户
• 基础设施依赖和中断
o 水
o 电力
o 通信（电话和互联网）
o 公路和船只
• 供应商
• 灾害准备
• 近期灾害的影响（详见表2）
• 疾病爆发的影响（寨卡和COVID-19）
• 财务账户

数据清理操作
---------------------------
为了匿名化目的，进行了以下数据编辑：
• 删除了精确的位置数据，如GPS坐标和次国家级别的行政区域（admin 1）
• 删除了识别和联系信息，如企业名称、受访者姓名、供应商名称、电话号码和电子邮件联系信息
• 将超过100名全职员工的数据重新编码为“超过100名全职员工”，以减轻对最大企业的重新识别。
有关匿名化的更多详细信息，请参阅技术说明。