Data
The record for each country’s international travel policy response to COVID-19 is obtained from the Oxford COVID-19 Government Response Tracker (OxCGRT) database [14]. The database records the level of strictness on international travel from 1 January 2020 to the present (continual updating), categorised into 5 levels: 0 - no restrictions; 1 - screening arrivals; 2 - quarantine arrivals from some or all regions; 3 - ban arrivals from some regions; and 4 - ban on all regions or total border closure. At various points in time from the beginning of 2020 to the time of writing, 73 countries have introduced screening on arrival policy, 77 have introduced arrival quarantine, 133 have introduced travel bans, and 137 have introduced total border closures.
Covid-19 statistics were obtained from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University [15]. The dataset consists of records on number of confirmed cases and death on a daily basis for 205 countries since January 2020.
Our measure of globalisation is generated from the 2019 KOF Globalisation Index (of more than 200 countries), published by the KOF Swiss Economic Institute[1] . The KOF Globalisation Index is made up of 44 individual variables (24 de facto and 20 de jure variables) relating to globalisation across economic, social, and political factors[2] (see also [12, 16]. The complete index is calculated as the average of the de facto and the de jure globalisation indices. In this analysis, we focus on the overall index as well as the major sub-components (i.e. Economic (Trade and Financial), Social (Interpersonal, Informational, and Cultural), and Political globalisation). Each index ranges from 1 to 100 (highest globalisation).
We also take into account that a country’s decision to adopt travel restrictions can be affected by the decision made by its (economic) neighbours. We constructed a variable to reflect this by averaging the international travel strictness of each country’s ‘neighbours’ weighted by share of international tourism and foreign trade. We obtained inbound tourism data of 197 countries from the Yearbook of Tourism Statistics of the World Tourism Organization [17]. The data consist of total arrivals of non-resident tourists or visitors at national borders or in hotels or other types of accommodations and overnight stays of tourists, broken down by nationality or country of residence, from 1995 to 2018. Due to difference in statistical availability for each country, we take the year 2018 record (or 2017 if 2018 is not available) of arrivals of non-resident tourists/visitors at national borders as the country weights for the computation of foreign international travel policy. If the records of arrivals at national border are not available for these years, we check for the 2018 or 2017 records on arrivals or overnight stays in hotels or other types of accommodation before relying on records from earlier years. To calculate the weighted foreign international restriction policy, for each country, we calculated the weighted sum using the share of arrivals of other countries multiplied by the corresponding policy value ranging from 0 to 4[3].
Additionally, we check our results using the share of total gross bilateral export or import in 2018 as the weights for constructing the weighted foreign policy variable. The data on trade, broken down by country, was obtained from the World Integrated Trade Solution – World Bank under the UN COMTRADE Standard International Trade Classification, Revision 4 (SITC Rev4) 2018 [18].
For additional control variables, we account for each country’s macroeconomic conditions, political, and geographical characteristics. First, we consider the country’s economic risk assessments taken from the International Country Risk Guide (ICRG), which is a composite rating accounting for factors such as inflation rate, real GDP growth, per capita GDP, balance of payment and current account as a percentage of GDP. From the World Development Indicators, we obtained the latest record of population density and the number of physicians per 1,000 people in the population, which we used to proxy for a country’s health system capability[4]. We also use the Boix-Miller-Rosato (BMR) dichotomous variable to identify democratic and autocratic countries[5] [19]. Lastly, we include continent dummies, whether the country is landlocked, and the land area (in log sq. km), which were obtained from GeoDist (CEPII) [20].
Empirical strategy
We hypothesize that more globalised countries are more likely to impose international travel restrictions later than less globalised countries. To test this hypothesis, we use records from the Oxford COVID-19 Government Response Tracker (OxCGRT; [14]) on the timing of restrictions on international travel for each country, COVID-19 case statistics from the Johns Hopkins University Center for Systems Science and Engineering COVID-19 dataset to derive our main dependent variable, namely, the time gap between the first national confirmed case and the first international travel restriction policy was implemented. We also calculate the number of days between the first confirmed case and each level of restriction imposed and test for the robustness of the results. Furthermore, we also conjecture that, as a consequence from the above, countries with higher levels of globalisation should have more confirmed cases by the time the first policy was introduced. Therefore, we also examine the relationship between globalisation and the number of confirmed case (in logs) at the time of policy implementation.
To study relationships between our outcome variables and the level of globalisation, we first present the simple correlations between them. We then apply ordinary least squares (OLS) regression models to estimate the following model:
where Yi is the number of days passed since the first Covid-19 case in country i to the implementation of travel restriction or the number of cases (in log) at the time of the restriction was implemented. is the KOF globalisation index of country i and X is a vector of country-specific controls such as the country’s health care capacity, economic risk, population density, geographical characteristics and the number of cases per million people in the population at the time of policy implementation.
Additionally, we examine how a country responds to the international travel policy implemented by those countries that contribute most towards its tourism sector. To do so, for each country, we constructed a variable based on the average strictness of international travel policy weighted by the share of tourists to the country of interest, calculated daily. We therefore include this variable, measured at the time of the focal country’s first implementation of the international travel policy into the regression.
[1] https://kof.ethz.ch/en/forecasts-and-indicators/indicators/kof-globalisation-index.html
[2] See https://ethz.ch/content/dam/ethz/special-interest/dual/kof-dam/documents/Globalization/2019/KOFGI_2019_method.pdf for detailed methods on the calculation of the weights of each component and the overall index