The rapid and global spread of the novel coronavirus (Dong et al. 2020) within just a few weeks is also infecting the global economy. Efforts to contain and mitigate the spread of the virus in order to reduce the loss of human lives have ultimate priority. Nevertheless, policymakers need to act quickly to address the mounting negative economic fallout from this crisis. To alleviate the negative medium- to long-term economic consequences, timely understanding of the effects on the consumer demand side of the economy is key to initiate the appropriate responses to stabilise the economy. A major worry is that a rise in consumers’ income and employment risk weakens their economic stability and their economic sentiment, fuelling a long-term economic downturn through an expectation-driven downward demand shock.
In a new paper (Fetzer et al. 2020), we document a rapid increase in economic anxiety during and after the initial global spreading of the novel coronavirus, inferred through daily Google search activity and individual survey data. Further, we shed light on how perceptions of the coronavirus shape economic anxiety and study the role of information and the underlying psychological mechanisms.
Documenting the rise in economic anxieties with Google searches and survey data
We first leverage a global dataset of Google searches from 190 countries to document a rise in economic anxieties during the initial global spread up to 29 February. After the arrival of the first confirmed case of coronavirus in a country, we observe a notable and sharp increase in Google search activity around topics such as recessions, stock market crashes, survivalism and conspiracy theories. This type of search activity increases sharply by between 20-50% relative to the period before the first coronavirus case arrived in a country.
We validate that Google searches are an important leading indicator of subsequent economic slowdowns and recessions. Using a historical panel leveraging quarterly data on real growth, we observe that these sharp increases in search activity, in particular for the term “recession”, are associated with subsequent marked economic slowdowns particularly due to lower real growth rates in consumption. We also document a further strong increase in economic anxiety after the arrival of the coronavirus measured as directly through nationally representative survey data for the US. These data were collected in a first wave on 5 March and in a second wave eleven days later, on 16 March. In the early March survey, the US showed 176 confirmed cases of the new coronavirus. By 16 March, this number had shot up to 4,576 cases – an increase by a factor of 26. The rapid rise in infections coincided with a sharp increase in perceptions of the severity of the coronavirus crisis. While on 5 March, 55% of respondents agreed that the US would be severely affected by the coronavirus, 78% of respondents agreed with this statement on 16 March.
Figure 1 Economic anxiety over time
Notes: The figure shows the evolution of beliefs about the severity of the crisis and economic worries between early- and mid-March.
Source: Fetzer et al. (2020).
Moreover, economic worries due to the coronavirus increased sharply: on 5 March, 68% of respondents were worried or very worried about the effects of the virus on the US economy. By 16 March, that share had increased to 88%.
Importantly, respondents’ concerns about their own personal economic situation also grew sharply, from 47% to 74% within just 11 days. As economic expectations are an important driver of consumption demand (Carroll et al. 1994, Roth and Wohlfart 2020), this suggests that, irrespective of the many direct adjustments to consumer behaviour that will be due to the containment measures, consumer demand will drop, compounding the negative economic effects.
How do perceptions of the new coronavirus shape economic anxiety?
To shed some light on the role of information and underlying psychological mechanisms that link economic anxiety to perceptions of the coronavirus, we elicited people’s beliefs about two salient characteristics of the new coronavirus: case mortality and its contagiousness (R0). We compare these beliefs to estimates that the medical literature documented so far.
Empirically, individuals’ beliefs about the mortality of the coronavirus are quite heterogeneous. More than 50% of respondents think the mortality is higher than 5% (average at 19%), which is much higher than most estimates in the medical literature as well as the current official WHO estimate of 3.4%. Beliefs about the contagiousness are also heterogeneous, with 50% of respondents believing that every infected person infects at least ten others. In comparison, initial medical research suggested an R0 of 2–3 (Li et al. 2020, Wu et al. 2020).
Figure 2 Perceptions of coronavirus
Note: The figure plots the distribution of beliefs about mortality and contagiousness (R0) of the novel coronavirus as measured in a representative sample (N=915) of the US population on 5 March 2020.
Source: Fetzer et al. (2020).
We further investigated the association of these beliefs with economic anxietiy. People who hold beliefs about the contagiousness and mortality of coronavirus that are higher than the scientific and official estimates also exhibit significantly higher degrees of anxiety about the disease. This suggests that more negative beliefs about the coronavirus are crucially shaping economic worries during this crisis.
This may not necessarily be a bad thing. If such anxieties and beliefs change the way people behave and interact, it may be effective in mitigating the spread of the disease. However, it could also increase anxieties further and exacerbate the negative economic fallout due to likely downward adjustment in consumption behaviour.
Mental models of disease spread
While respondents appear to overestimate the mortality and contagiousness of the coronavirus compared to official estimates, they appear to significantly underestimate the non-linear nature of infectious disease spread. In the second survey experiment on 16 March, we investigated whether respondents adequately capture non-linear growth patterns by asking them to predict the spread of a fictitious disease. Participants were instructed to assume that the case number of the fictitious disease on day 1 is equal to one, and each day a newly infected person infects two healthy people and then stops being contagious.
Individuals were found to highly underestimate the spread of the fictitious disease. In contrast to correct prediction values of 31 on day 5, 1,023 on day 10, and 1,048,575 on day 20, the median participant estimated a case number of 16 on day 5, 30 on day 10, and 60 on day 20. The predictions of the median participant could be well approximated by a linear mental model.
Figure 3 Mental models of disease spread
Note: This figure displays participants’ median, 75th percentile, and 90th percentile belief about the spread of a fictitious disease on a logarithmic scale. Participants were instructed to predict the number of cases of a fictitious disease on day 5, 10, and 20. Participants were informed that on day 1, one person has the disease and that each day a newly infected person infects two healthy people and then stops being contagious. In both panels, the blue line indicates the correct prediction, the green line an incorrect linear model with a growth rate of two per day.
Source: Fetzer et al. (2020)
Furthermore, we find a strong and positive association between people’s estimates of the number of infected individuals and their worries about the US economy. These results indicate that individuals who exhibit a more accurate mental model of non-linear growth in disease spread also exhibit higher worries regarding the coronavirus pandemic, potentially as they foresee a greater potential for a widespread contagion of the global population.
Communication and education can shape economic anxieties
A natural question for governments and decision makers across the globe is how the economic fallout from the current crisis can be contained. The media may be playing an important role in shaping how people perceive the underlying risks. Through an experiment embedded in our survey, we randomly exposed a subset of the respondents to information comparing the estimated mortality of the coronavirus with that of the regular seasonal flu. Another subset of people were given a comparison of the mortality estimates with that of SARS.
We find that comparing the mortality of the coronavirus with that of the flu increases reported economic worries. This highlights that the way the press conveys information about the coronavirus may matter a lot in shaping the extent of economic anxieties. Given the large heterogeneity in beliefs about key characteristics of the coronavirus, our findings suggest that information and public education may play a central role in achieving successful containment and managing the negative economic impact of increased economic anxiety.
Policy interventions to alleviate economic hardship and anxieties
Our evidence highlights a rapid increase in economic anxiety in the population at large. Because at this time a surge in unemployment numbers across several countries can be expected, measures that directly reduce economic hardship and anxiety will be needed. While counter-cyclical policies such as government spending can help to stabilise the demand side in the medium to long run (Auerbach and Gorodnichenko 2012), a more direct and short-term tool to prevent immediate economic instability involves cash transfers, as have been currently initiated by several governments. Cash transfers will help economically vulnerable populations to cover expenses, alleviate economic hardship in the short term, and stabilise economic demand. Moreover, cash transfers have been shown to reduce psychological distress and anxieties (Haushofer and Shapiro 2016, Christian et al. 2019).
Auerbach, A J and Y Gorodnichenko (2012), “Measuring the output responses to fiscal policy”, American Economic Journal: Economic Policy 4(2): 1-27.
Christian, C, L Hensel and C Roth (2019), “Income shocks and suicides: Causal evidence from Indonesia”, Review of Economics and Statistics 101(5): 905-920.
Carroll, C D, C Fuhrer and D W Wilcox (1994), “Does consumer sentiment forecast household spending? If so, why?”, The American Economic Review 84(5): 1397–1408.
Ensheng, D, D Hongru and L Gardner (2020), “An interactive web-based dash- board to track covid-19 in real time”, The Lancet Infectious Diseases.
Fetzer, T, L Hensel, J Hermle and C Roth (2020), “Coronavirus Perceptions and Economic Anxiety”, arXiv preprint arXiv:2003.03848.
Haushofer, J and J Shapiro (2016), “The short-term impact of unconditional cash transfers to the poor: Experimental evidence from Kenya”, The Quarterly Journal of Economics 131(4): 1973-2042.
Li, Q, X Guan, P Wu et al. (2020), “Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus–Infected Pneumonia”, The New England Journal of Medicine.
Roth, C and J Wohlfart (2020), “How do expectations about the macroeconomy affect personal expectations and behavior?”, Review of Economics and Statistics, forthcoming.
Wu, K, K Leung and G M Leung (2020), “Nowcasting and forecasting the potential domestic and international spread of the 2019-nCoV outbreak originating in Wuhan, China: a modelling study”, The Lancet 395(10225): 689-697.