COVID-19 and tourism: What can we learn from the past?

Abstract

The impact of the COVID-19 crisis on tourism flows is without precedent in terms of speed and severity. In this paper, we try to infer a possible future scenario for the tourism sector, evaluating the medium-term effects of past pandemics on tourist arrivals. We find that pandemics lead to a persistent decline in tourist arrivals, with the effects being larger in developing and emerging countries. Interestingly, the effects are heterogeneous across countries and episodes, and depend on several economic conditions such as the overall health system performance, the severity of the shock, and the uncertainty induced by the pandemic event.

Keywords: COVID‐19; health systems; international arrivals; pandemics; tourism; uncertainty.

FIGURE 1

Impact of pandemics on tourist arrivals (%). Note: The chart shows the impulse response functions and the associated 90 per cent confidence bands; t = 0 is the year of the pandemic event. Estimates based on equation (1) using a sample of 183 countries over the period 1995–2018

FIGURE 2

Impact of pandemics on international tourist arrivals (%) – Robustness checks. Note: Impulse response functions are estimated using a sample of 183 countries over the period 1995–2018. The graphs show the response and 90 per cent confidence bands. The x‐axis shows years (k) after pandemic events; t = 0 is the year of the pandemic event. The specification in Panel (b) includes several control variables such as proxies for the level of economic development (log of real GDP), trade openness (imports and exports as a share of GDP), international competitiveness (price level ratio of PPP conversion factor (GDP) to market exchange rate) and population density (people per sq. km of land area).

FIGURE 3

Impact of pandemics on tourist arrivals (%) – by country groups. Note: Impulse response functions are estimated using, in turn, a sample of selected countries (38 AE; 90 EME; 55 LIDCs) over the period 1995–2018. The chart shows the response and 90 per cent confidence bands; t = 0 is the year of the pandemic event. Estimates based on equation (1)

FIGURE 4

Impact of pandemics on tourist arrivals (%) – by Health System Performance. Note: Impulse response functions are estimated using a sample of 183 countries over the period 1995–2018. The graphs show the response and 90 per cent confidence bands. The x‐axis shows years (k) after pandemic events; t = 0 is the year of the pandemic event

FIGURE 5

Impact of pandemics on tourist arrivals (%) – by pandemics. Note: The charts show the impulse response functions and the associated 90 per cent confidence bands; t = 0 is the year of the pandemic event. Estimates based on equation (1) using a sample of 183 countries over the period 1995–2018

FIGURE 6

Impact of pandemics on tourist arrivals (%) – The role of the number of cases. Note: Impulse response functions are estimated using a sample of 183 countries over the period 1995–2018. The graph shows the response and 90 per cent confidence bands. The x‐axis shows years (k) after pandemic events; t = 0 is the year of the pandemic event. Estimates based on equation (2) using the ratio of the number of cases to population as a state variable. The dotted blue line denotes the average (unconditional) effect reported in Figure 1. The red lines indicate the estimates for pandemic events associated with very low and high ratio of the number of cases to population

FIGURE 7

Impact of pandemics on tourist arrivals (%) – The role of uncertainty. Note: The charts show the impulse response functions and the associated 90 per cent confidence bands; t = 0 is the year of the pandemic event. Estimates based on equation (2) using a sample of 183 countries over the period 1995–2018. The dotted blue lines denote the average (unconditional) effect reported in Figure 1. The red lines indicate the estimates for pandemic events associated with very low (left panel) and high (right panel) uncertainty