The perils of trying to distil big picture lessons about wars in the midst of ‘the fog of war’ are well known. The same risk surely extends to the COVID-19 global pandemic. Yet strategic planning and revising plans on-the-fly is necessarily real-time, pricing assets is real-time, and arguably some tentative parts of the big picture are also already starting to emerge.
Differences in cross-country approaches could potentially be particularly instructive, yet frequently devolve into ideological and political debate. For example, the shorthand has become: “Are we on the Italy trajectory or the Singapore /South Korea/Taiwan/Hong Kong – perhaps also Japan – trajectory?”. The latter group, in particular Singapore, has been dubbed the gold standard.1 It is true that the visible hand in these gold standard countries facilitated responses to the crisis for which they were well-prepared with rapid testing and contact tracing. Meanwhile, countries like the US and Australia tend to be relatively more decentralised when the ‘rubber hits the road’ in terms of planning and provision for emergency services. They are provided at the municipal and state levels – mutual aid networks help insure against unexpectedly severe local or regional catastrophes like hurricanes, fires, earthquakes or influenza overloads. But shuttling mutual aid resources between municipalities or states is not a solution when there’s a common factor causing emergencies across virtually all districts at the same time. Asia had also prepped for the common viral calamity via their SARS experience nearly two decades ago.
However, cross country generalisations about disaster recovery planning are also rarely black and white. On the one hand, the apparently decentralised US has a military with centralised nationwide defence network planning; on the other hand, in centralised China, municipal authorities were alleged to have tried to thwart early signals of the novel virus – though only with hindsight could any authority understand how major the implications would turn out to be.2
Crisis planning also involves a well-known political/agency problem: it’s hard to see the costs of infrequent crises that are averted because of good planning, but it’s easy to see the up-front planning and preparation costs. In the same way that securities portfolios with a cash allocation provide downside risk protection that looks expensive in a bull market, so too do resources tied up in hospital beds sitting empty, critical personnel planning, and full warehouses, in order to cover peak demand when no virus attacks. Strong leadership is needed to keep funding the ‘insurance premiums’!
Given some initial level of insurance/preparation, most observers argue that both sickness and healthcare costs on one side, and economic costs of lockdowns/shutdowns (‘collateral damage’) on the other, are smallest when the outbreak is addressed quickly with extensive testing, contact tracing, and isolation/quarantining (the ‘Singapore model’). At later stages, targeted testing and contact tracing are well-nigh infeasible given by-then widespread transmission. The lockdowns in place now in countries like Australia, the US, Italy, and so on are a sledgehammer ‘flattening’ of the much-cited curve for required medical resources – the sledgehammer results in substantial economic collateral damage.
We can do some back-of-the-envelope calculations. Barro (2006) calibrated a model for ex ante pricing of risk of major economic disasters in the 20th century defined, for a separate country, as those resulting in a 15% or greater short-run drop in GDP. (Critics of the lockdowns will quip that the lockdowns’ man-made effect on business and production will result in the coronavirus crisis easily hitting the 15% threshold.) The ex ante insurance premium for holding a market portfolio of corporate equity exposures to disasters is part of the average equity premium. Barro argues that the 20th century average market premium calibrates with a 1% annual probability of a global disaster event that causes a one-third to one-half short-run GDP drop, which is in turn roughly the record of the 20th century. Figure 1 shows the stock market (the Dow Jones Industrial Average or DJIA) plotted from five days prior to six prominent historical crises to 12 months thereafter.
Note: In the above figure, we assume February 21, 2020 (the day when Italy reported its first death) as “day 0” for the COVID-19. To facilitate comparison, the DJIA price index for each line is normalized to $1 on day –5.
Note that the US market has eventually recovered from all crises: we are looking at a market which survived. That said, along the way the market crashes and recoveries gave a whole new extreme sports meaning to the term ‘buying on the dips’ (or ‘holding on the dips’) in the worst days. Marsh and Pfleiderer’s (2013) calibrations for asset allocation in the 2007-2009 financial crisis suggest annualised ex ante equity premiums in the vicinity of 30% to 35% were priced into equities in order to clear the market in the throes of the crises. Two of the crises, 9-11 and the Long-Term Capital Management (LTCM) crisis, were characterised by quite rapid recoveries, again not surprising since the 9-11 event proved to be a largely ‘one time only’ terrorist event, while the LTCM market nose-dive reflected market illiquidity following the Russian sovereign default.
The Dow’s trajectory in the coronavirus crisis has more resembled that of the Great Depression and the 2007-2009 subprime crisis – given the US’s latest economic numbers,3 the resemblance appears to reflect rational market anticipation of the business consequences, not the long-run costs of the virus itself. The fall in equity prices since 21 February is also much bigger than in previous infectious disease cases, including the 1918 Spanish influenza. This is again not surprising since none of the previous infection cases involved bringing production to a virtual standstill – this is shown dramatically in Figure 2.
The evidence is that equity market prices like the Dow Index are negatively impacted by increases in uncertainty – increases in market uncertainty over time make large negative price changes more likely! Sources of uncertainty in the coronavirus case could be parsed into:
(1) the unknown novel technical properties of the novel virus;
(2) the business ramifications of regulatory policies, subsidies etc. put in place to contain the spread of the virus; and
(3) ‘new normal’ changes in global supply chains and the like that might be wrought by the far-reaching crisis (we omit discussion of (3) here).
Interestingly, (1) is being resolved thanks to dedicated 24-7 work in biomedical research. Expert opinion suggests that the technical virus uncertainty is likely to be resolved in the roughly one-year-or-so horizon that it will take for a vaccine and treatment to be developed – significantly, this horizon spans includes the 2020 presidential election and Northern Hemisphere summer and winter seasons. The resolution of uncertainty would be good news for the securities markets, but the trajectory over the year is unlikely to be monotonic – unlike, say, the resolution of economic uncertainty associated with national US elections, which we documented in Chan and Marsh (2017) and which extends over roughly six months. Consider the ‘second wave’ planning in a country like Singapore: its initial success in containment can only be maintained by controlling second and subsequent waves of new ‘imported’ cases. The economic fallout from controlling this second wave is likely to pose major economic frictions. Imagine the ‘Singapore second wave control problem’ on a bigger scale – say, across countries in the EU or even between hot-spot and ‘cool-spot’ regions of countries (“Western Australian visitors, please present your negative-test card upon arrival at Perth airport?”).
Item (2) arguably poses the greatest uncertainty over the next year. As this week’s Economist cover has it, the Earth is largely “closed for business.” Baker, Bloom and Davis’ updated Economic Policy Uncertainty (EPU) index is at an all-time high. Depending on the complexity of how economies ‘power back up’, it is also conceivable that the lockdown exit uncertainty will be accentuated by a cascade of corporate and consumer bankruptcies (given the substantial US consumer debt outstanding) – A can’t pay B until C pays A, precisely as in a case where a common global factor shock hits economic clusters simultaneously. Note that A could well be a state or government agency or municipality. As we write this, the US has just announced a $2 trillion ‘stimulus package’, though it seems more likely that the $2 trillion will be plugging the already-induced fiscal shortfalls than be an incremental stimulus. The topic of conversation here may soon move from the silent corona virus to the silent inflation tax.
Barro, R (2006), “Rare Disasters and Asset Markets in the Twentieth Century”, Quarterly Journal of Economics 121: 823-866.
Chan, K F and T Marsh (2017), "Equity Premiums in the Presidential Cycle: The Midterm Election Resolution of Uncertainty”.
Marsh, T and P Pfleiderer (2013), “Flight to Quality and Asset Allocation in a Financial Crisis,” Financial Analysts Journal 69: 43-57.
2 Also, the planning and execution in Wuhan still looks swift by comparison with much Western government action two months later!
3 For example, jobless claims were announced on Thursday 26 March to be 3.3 million, roughly five times their previous record high.