DQ-en

Ever stricter policy interventions lower COVID-19 infections and deaths but degrade economic performance through introducing frictions into social activity. Thus, the correlation should be negative between measures of economic performance and anti-COVID social protection. Cross-country evidence, however, shows the value of that correlation to be zero. This article considers the hypothesis that anti-pandemic policies have different, opposing effects on economic activity. The relative size of those different impacts both explain the zero cross-country correlation and provide a narrative for observed varied national success. Which effect dominates is crucial for determining the right anti-pandemic policies for specific national circumstances.

The Facts to Uncover and Their Policy Significance

Perhaps the single most widespread and memorable economic picture of the global COVID-19 pandemic is a graph such as Figure 1. This graph shows coordinated 2020 downturn in growth across economies around the world. (Names of the economies corresponding to the ISO3c codes here appear in the Table below.) The pandemic ushered in synchronized global economic contraction.

However, Figure 1 conflates two effects. First, when policymakers combat the pandemic by imposing lockdowns, they reduce valuable economic transactions. Second, regardless of lockdown, economic activity is already curtailed by the pandemic fallout of illness, mortality, loss of consumer and investor confidence, and heightened risk aversion. In Figure 1 both effects operate in the same direction to worsen economic downturn.

For policy, however, the two effects work in opposite directions. The first effect implies that when policymakers continue lockdown policies they will further shrink the economy. The second effect implies that those same restrictive policies promote economic growth because they reduce illness and mortality, and restore confidence. Depending on which effect dominates, identical policy actions can either contract or expand the economy. Policymakers need to assess the relative strength of the two effects to best combat the pandemic.

A univariate analysis of a kind surrounding Figure 1 cannot disentangle these two different effects. A second indicator needs to be brought into the frame.

Two Variables to Uncover Two Underlying Mechanisms, But Even That is Not Enough

The logic is straightforward that policy interventions that seek to break the chain of COVID infection, and thus reduce COVID deaths, also lower economic activity. Frictions in value creation result from restricting operations for bars, restaurants, and other entertainment outlets; curtailing vacation and business travel; or in general making mobility more stringent. Such disruptions destroy jobs and businesses. Longer term, human and social capital will deteriorate through school closures and reduced business and personal contact.

In this reasoning, there is a tradeoff between social safety and economic performance. If a society seeks more of one, it ends up with less of the other. Raising stringency measures for improved social safety will generate frictions in economic activity, and thus reduce economic prosperity. Call this the stringency effect. Data should show a negative relationship between social safety and economic growth because of the stringency effect. The dilemma for policy-making is a choice between COVID safety and economic prosperity, with both valued by societies, but one coming at a cost to the other.

In reality, however, this tradeoff is offset by an opposing force. Even absent stringency effects, COVID-19 is simultaneously a negative supply shock and a negative demand shock. From the pandemic, all else equal, output will fall from ill health and mortality in the workforce. With heightened public concern over COVID, consumption will decline for a combination of reasons: heightened job insecurity and lowered consumer confidence; increased savings; lowered propensity to take vacations, eat restaurant meals, or purchase leisurewear. Increased risk aversion will depress entrepreneurial activity and reduce investment. In this reasoning, combating the pandemic helps lift economic activity. .The higher is social safety, the greater is confidence that life will return to normal, and thus the higher will be economic growth. The correlation here is positive between social protection and economic prosperity. Call this the shock effect.

A further complication over the longer run is that the economy will experience reallocation effects due to efforts to combat COVID: restrictions on mobility and physical clustering harm the travel industry but induce growth and employment in sectors such as those that develop tools for digital engagement, distance working, and telemedicine. As this Schumpeterian creative destruction proceeds, the same stringency conditions that initially lower economic activity can, conversely, spur innovation and growth.

Simply looking at correlations between COVID-19 and economic growth conflates stringency and shock effects as these latter exert opposing pressures. A more nuanced econometric analysis is needed to disentangle the underlying information in the data.

Empirical Findings: Unpacking the Correlation

Across all nations, the estimated correlation between COVID safety and economic prosperity is positive but not significantly different from zero. Neither the stringency nor shock effect is dominant.

Specifically, measuring COVID-19 performance by the reciprocal of accumulated deaths per million, and economic performance by the difference between economic growth rates over 2020-2021 and 2016-2019, then the raw Pearson correlation between COVID-19 performance and economic performance is only 0.09. The 95% confidence interval is (-0.7, 0.25), with marginal significance level 26% against the hypothesis that the correlation is 0.

My COVID-19 performance measure assesses how well policy measures and social behaviour have been able to keep COVID-19 mortality low. The higher is this number, the more successful has that society combated COVID-19.

I have chosen my measure of COVID-19 safety to be mortality in the population (or its reciprocal), rather than the setting policymakers select for their national stringency measures. This is because those last are only instruments whose effectiveness depend on population compliance, social attitudes, communication credibility and effectiveness, and so on, that lie outside the direct control of the policymaker. Whether stringency measures are set high or low translate only imperfectly to the results that ultimately matter to the representative agent.

My economic performance measure quantifies both the size and direction of the economic shock over the COVID-19 period. The higher algebraically is this measure of economic performance, the more successful has that economy been. This measure takes into account how certain economies have historically been fast-growing. Thus, that China’s economic growth over this pandemic period has remained high does not by itself imply economic success. The important question is, how well did China do compared to how it was doing before the pandemic hit?

But even with these, simply focusing on a single correlation shoehorns the analysis into a linear, one-size-fits-all framework that turns out to be misleading.

Figures 2 and 3 provide granular information on the relation between COVID-19 and economic performance across nations. The Figures differ in the variable that appears on the horizontal axis: Figure 2 places Economic Performance there; Figure 3, COVID safety.

As expected, since the overall correlation is positive but not significantly different from zero, Figures 2 and 3 will have best-fitting nonlinear regression lines that are on average approximately horizontal but with a slight positive slope. That nonparametrically-estimated regression line should be viewed as a state-dependent projection of COVID and economic outcomes onto whichever variable appears on the horizontal axis.

Figure 2, which places Economic performance on the horizontal axis, shows on its leftmost portion a negative slope. But this section of Figure 2 is the only region in both Figures where the stringency effect manifests,. It is only here that there is a visible tradeoff between COVID and economic performance.

Everywhere else in both Figures, the shock effect dominates, where confidence in successfully combating the pandemic effect comes with improved economic performance. Thus, in by far the greater fraction of the space of possible outcomes, ensuring public safety lifts, rather than compromises, economic performance.

How Have Individual Nations Fared?

In Figures 2 and 3 the respective scatterplots and best-fitting nonlinear regression lines also show how individual nations have succeeded or failed relative to average.

China appears in the upper portion of both Figures relative to the regression curve. This means that, conditional on its economic performance, China has been more successful than average in preserving social safety. At the same time , conditional on its COVID performance, China has been more successful than average in keeping its economy going.

The US appears in the upper portion in Figure 3, where COVID is on the horizontal axis. Thus, given its COVID performance, the US has kept successful economic growth. However, that the US appears in the lower portion in Figure 2 means it has suffered unexpectedly high COVID deaths, given its economic performance.

The UK appears below the regression line in both Figures 2 and 3: It has experienced both below-average pandemic safety and economic performance. Failure has occurred on both fronts.

Like China, Singapore appears in the upper portion of Figure 2: its pandemic safety performance has been above average. In Figure 3, however, Singapore shows up just below the regression line: This is consistent with Singapore’s policy-makers insisting that its economic management was intended to keep the economy from excessive recession but, from an abundance of caution, not inducing unnecessary economic expansion that might potentially jeopardise public safety.

Circumstances continue to change, obviously, with the Delta variant of the coronavirus still working its way through the global system. However, to date, both India and Indonesia are approximately on the regression line for both Figures. Cambodia has seen successful COVID safety but with economic performance below average. The Philippines has kept to an average COVID safety but with a dismal economic outcome. Myanmar appears well to the left in Figure 2 reflecting the double effect of both the COVID shock and the ongoing military coup.

Thailand and Vietnam have had similar experiences. Both appear in the upper portion of Figure 2 meaning they show above average pandemic safety conditional on economic performance. However, they are both also in the lower part of Figure 3 meaning that conditional on COVID safety performance, they show below-average economic results.

Figures 2 and 3 can be put together on a single graph, Figure 4, that provides in this work a counterpart to the classical phase diagram for analysing dynamics. The two loci in blue separate below and above average performance estimated by state-dependent projection onto the respective axes. The extreme northeast quadrant show those economies that have achieved above-average performance both in COVID safety and economic prosperity; the extreme southwest, the opposite. Further analyses using this quasi phase diagram approach will be reported in future work by the author.

Conclusion

Considerable policy discussion has focused on the stringency effect of policy interventions taken against the pandemic. Typically ignored in such debates is the shock effect, where combating the pandemic is good for the economy.

The raw evidence shows a near-zero correlation between COVID safety and economic performance outcomes. Large portions of the outcome space however indicate a positive rather than negative relation. In these latter regions, the shock effect dominates. Here, coming down hard on the pandemic improves economic performance.

A reasonable conjecture is that stringency effects are large in the short term, but shock effects emerge more gradually over time. In this view, the faster and harder the clampdown, the more likely will the stringency effect be done with, and the shock effect then surface.

Statistical Annex

The Figures use ISO3c designations for the different economies, instead of more familiar, human-friendly names. The charts become too cluttered otherwise. For convenience, however, the correspondence together with the calculated statistics are provided in the Table.

Economy	ISO3c	Covid	Economy
Brazil	BRA	0.45	-0.5
China	CHN	310.0	-1.2
Germany	DEU	0.94	-2.3
United Kingdom	GBR	0.53	-3.8
Indonesia	IDN	5.32	-4.0
India	IND	3.98	-4.1
Cambodia	KHM	66.3	-6.80
Myanmar	MMR	16.9	-9.18
Malaysia	MYS	9.83	-4.38
Philippines	PHL	5.04	-7.93
Singapore	SGP	177.0	-3.27
Thailand	THA	57.5	-5.28
United States	USA	0.55	-0.86
Vietnam	VNM	1838.0	-2.23

The Figures and Table exclude nations such as Lebanon, Libya, and Venezuela whose economic growth rates have been volatile for other, long-standing reasons. Similarly, the sample excludes those nations whose accumulated COVID deaths per million were lower than 0.5, as numbers so small are especially unstable in daily updates. The resulting sample turned out to exclude, among ASEAN’s member states, Laos but also, more generally, a number of other nations, including Burundi.

An edited and abbreviated version of this article appears as

Quah, Danny. 2021. “The Complicated Question of COVID-19 vs The Economy“, Asia Economics Blog (09 June) http://acaes.us/blog/covid-19-vs-the-economy