After-Effects of the COVID-19 Pandemic: Prospects for Medium-Term Economic Damage

WP/21/203

After-Effects of the COVID-19 Pandemic: Prospects for Medium-Term Economic Damage

by Philip Barrett, Sonali Das, Giacomo Magistretti, Evgenia Pugacheva, Philippe Wingender

IMF Working Papers describe research in progress by the author(s) and are published to elicit comments and to encourage debate. The views expressed in IMF Working Papers are those of the author(s) and do not necessarily represent the views of the IMF, its Executive Board, or IMF management.

© 2021 International Monetary Fund

WP/21/203

IMF Working Paper

Research Department

After-Effects of the COVID-19 Pandemic: Prospects for Medium-Term Economic

Damage1

Prepared by Philip Barrett, Sonali Das, Giacomo Magistretti,

Evgenia Pugacheva, Philippe Wingender2

Authorized for distribution by Malhar Nabar

July 2021

IMF Working Papers describe research in progress by the author(s) and are published to elicit comments and to encourage debate. The views expressed in IMF Working Papers are those of the author(s) and do not necessarily represent the views of the IMF, its Executive Board, or IMF management.

Abstract

The COVID-19 pandemic has led to a severe global recession with differential impacts within and across countries. This paper examines the possible persistent effects (scarring) of the pandemic on the economy and the channels through which they may occur. History suggests that deep recessions often leave long-lived scars, particularly to productivity. Importantly, financial instabilities-typically associated with worse scarring-have been largely avoided in the current crisis so far. While medium-term output losses are anticipated to be lower than after the global financial crisis, they are still expected to be substantial. The degree of expected scarring varies across countries, depending on the structure of economies and the size of the policy response. Emerging market and developing economies are expected to suffer more scarring than advanced economies.

JEL Classification Numbers: E32, N10, O47

Keywords: COVID-19, scarring, medium-term output

1. We are grateful to John Bluedorn, Petya Koeva Brooks, Gita Gopinath, and Malhar Nabar for invaluable guidance and support , and to Weicheng Lian for helpful discussions. We thank Srijoni Banerjee, Savannah Newman, and Jungjin Lee for outstanding research s upport. Some of the analysis presented in this paper was published in Chapter 2 of the April 2021 World Economic Outlook, Internation al Monetary Fund.
2. All authors are at the International Monetary Fund. Authors' E-Mail Addresses: [e-mail],[e-mail],[e-mail],[e-mail],[e-mail].

2

I. Introduction

The COVID-19 pandemic has led to a severe global recession that is unique in many ways. The contraction in 2020 was very sudden and deep compared to previous global crises, even as the policy response in many countries was swift and sizable. The pandemic crisis also stands out for its differential impacts across sectors and countries, complex channels of transmission, and high uncertainty about the recovery path, given that it depends on the fate of the virus itself. The extent of scarring (persistent damage to supply potential)3 following the recession will differ across countries, as the health crisis interacts with countries' economic structures (such as the importance of "high-contact" sectors, where people are in close proximity) and varying policy responses.

The atypical features of the crisis-its severity, differential impacts, complex transmission, and high uncertainty-make assessment of the economic effects of COVID‑19 challenging. This paper aims to shed light on the potential main channels of scarring post-COVID-19 and implications for the medium-term outlook. We first ask what can we learn about prospects for scarring from historical experience with recessions? What are the most relevant channels in the current setting (productivity, labor, capital)? We draw lessons from previous recessions including those associated with past pandemics and epidemics, financial crises, natural disasters, and violent conflict outbreaks. Second, we investigate expectations of scarring, by comparing current forecasts for medium-term output with those from immediately before the onset of the pandemic. We further explore what factors-such as the income level, the sectoral structure of the economy (its precrisis dependence on tourism and its precrisis services share), and the size of the fiscal policy response in 2020-help explain the variation in expected medium-term outcomes (using a five-year horizon, i.e. by 2024) across economies.

Our findings from historical episodes suggest that severe recessions in the past have been associated with persistent output losses. The greatest scarring in the past has occurred in recessions associated with financial crises. Experience from previous recessions also suggests that the productivity channel could be particularly important, as these recessions have been followed by persistent losses to total factor productivity (TFP).

For the COVID-19 pandemic, expected medium-term output losses are sizable, but they exhibit significant variation across economies and regions . Despite higher-than-usual growth as the global economy recovers from the COVID-19 shock, world output is still anticipated to be about 3 percent lower in 2024 than pre-pandemic projections suggested (see IMF 2021a). This expected scarring is less than what was seen following the global financial crisis, consistent with the assumptions that financial sector disruptions remain contained in the recovery from the current crisis and that the pandemic is brought under control globally by the end of 2022. Unlike during the global financial crisis, when advanced economies were much more affected, emerging market and developing economies are expected to have deeper scars than advanced economies. This reflects in part their more muted policy responses, as countries with larger pandemic- related fiscal responses are projected to experience smaller losses. After accounting for income

3 Such supply damage could result from the loss of economic ties in production and distribution networks arising from job destruction and firm bankruptcies.

3

differences, economies that are more reliant on tourism, and those with larger service sectors, are projected to experience more persistent losses.

Our results are in line with previous literature, which suggests that output losses following recessions are persistent, particularly after financial crises, with differential impact across country groups. Cerra and Saxena (2008) find that currency crises lead to permanent output losses ten years after onset, with more adverse impacts for middle- and low-income countries, and that banking crises or concurrent twin crises have even more adverse effects. Moreover, Blanchard, Cerutti, and Summers (2015) find that recessions in general, and also those associated with financial crises and oil price increases, are often followed not only by lower output level, but also lower growth, implying that the scarring effect increases over time. Ball (2014) likewise points to significant scarring following the global financial crisis, with an adverse effect on output growth. Abiad and others (2009) and Chen, Mrkaic, and Nabar (2019) also document larger output losses following banking crises, stemming from lasting declines in capital per worker, TFP, and employment. Adler and others (2017) analyze the widespread decline in TFP growth following the global financial crisis and document that it has been persistent and was the main contributor to output losses relative to the precrisis trend.

There are also several recent studies that focus on the economic impact of past pandemics and epidemics. These include Jordà, Singh, and Taylor (2020), who find that macroeconomic effects of pandemics persist for decades, leading to a decline in real interest rates; Ma, Rogers, and Zhou (2020), who find that following the initial decline the bounce-back in output is rapid, but remains below pre-recession level five years after the shock; and Barro, Ursúa, and Weng (2020), who attempt to disentangle the effects of the Spanish flu and WWI deaths and find that GDP per capita declined by 6 percent as the result of the pandemic, which was on par with the 8.4 percent decline associated with the war.

Our main contributions to the literature on the economics effects of recessions are to conduct a comprehensive analysis of past recessions, using a broader sample of 586 recession episodes from 115 countries over 1957-2019, and to study the channels through which persistent damage occurs, by analyzing the effects of recessions on the supply-side components of GDP. In addition, we differentiate between deep and shallow recessions, as the impact of COVID -19 may be more like that of past recessions that likewise resulted in a large drop in output in the year of the impact, and differentiate between short recessions that last one year and longer recessions in which output declines for longer than a year, as some countries are expected to recover faster from COVID-19 than others. Like previous studies, we differentiate between different types of crises (past pandemics and epidemics, financial crises, natural disasters, and violent conflict outbreaks), but do it in a unified framework in which all types of recessions are analyzed within the same regression via interaction terms, which allows us to account for potential co-occurrence of several types of crisis events. We further contribute to the study of the macroeconomic effects of the COVID-19 pandemic by shedding light on the factors that explain differences in expected medium-term outcomes across countries.

The rest of the paper is organized as follows: Section II describes the data used in the analysis, Section III looks at the impact of past recessions on aggregate output and the channels

4

of impact, also differentiating recessions by their depth and duration, Section IV presents the expected medium-term output losses following the pandemic and analysis of factors that drive forecast revisions, and Section V concludes.

II. Data

The historical analysis relies on the Penn World Table (PWT) 10.0 database (Feenstra, Inklaar, and Timmer 2015), from which we draw on data for real GDP per capita (at constant prices in 2017 US dollars) that we use to identify recession episodes and to quantify the aggregate impact of those recession episodes on the economy. We also look at the supply-side channels of scarring (capital, labor, and productivity) using PWT data on capital stock (per person engaged), number of persons engaged (as employment-population ratio), and total factor productivity.

Recession episodes and the corresponding peaks and troughs of the cycle are identified using the Harding and Pagan (2002) algorithm on annual real GDP per capita, with a window of 1 year, minimum phase length of 1 year, and minimum cycle length of 2 years. While the standard approach for business cycle dating is typically done using quarterly data, the use of annual data allows for the identification of cycles for a larger sample of countries, in particular including developing economies for which quarterly data is often not available. Recessions identified using this approach for the United States match those reported by the NBER.

Recessions are further classified by co-occurrence of a particular type of a crisis, namely: a financial crisis, an epidemic or pandemic, a disaster, or a violent conflict. Each recession can be associated with several types of crises, or with no crisis, in which case it is referred to as a "typical" recession. The incidence of financial crises follows Laeven and Valencia (2018) for the period going back to 1970 and Reinhart and others (2016) for years prior to 1970. In both cases, financial crises include banking crises, currency crises, and sovereign debt crises. Past modern epidemics and pandemics include the Hong Kong flu, SARS, H1N1, MERS, Ebola and Zika and are identified for countries in which cases have been reported (Furceri and others 2020; Cockburn, Delon, Ferreira 1969). Disasters are identified using the Emergency Events Database (EM-DAT) when a country in a given year has experienced disasters that led to damages exceeding 1% of GDP or affected 5% of population (including deaths). Finally, a country is defined as being in conflict if in a given year there are battle-related deaths that exceed 100 people per one million population (Novta and Pugacheva 2021).

The analysis of expected medium-term output losses following the COVID-19 crisis, rely on the comparison of growth forecasts made by economists at the International Monetary Fund (IMF) presented in the World Economic Outlook (WEO) publication. These forecasts are available up to five years ahead for 194 countries. The forecasts are revised regularly and published twice a year (around April and October), with additional updates made in between the publications (around January and July). Data availability for a large number of countries and a record of forecast revisions at different stages of the pandemic thus make the IMF forecasts a good resource for analyzing the impact of the COVID-19 pandemic. Specifically, we measure medium-term output losses (or gains) as the difference between the level of real GDP forecast

5