Complex systems theory and cognitive biases are heuristically useful tools for thinking about the present coronavirus epidemic. Complex systems are commonly characterised by (1) sudden transitions and non-linearities, (2) limited predictability or fundamental uncertainty, (3) large events, (4) evolutionary dynamics (5) self-organisation and (6) emergence. In such systems, small, insignificantly seeming events can cause large events [(1) and (3)]. This is popularly known as the butterfly effect, whereby a small local change (in a deterministic, but non-linear) system can lead to large (non-linear) differences at a later stage (the butterfly flapping its wings in Brazil causes a hurricane in Canada – or something to that extent).
From a reverse angle the butterfly effects turns into a butterfly defect (Goldin & Mariathasan 2014). More concretely, globalization can be seen as having created a complex system with various interdependencies and connectivities. As such, the economy, logistics, health etc. susceptible to small events causing outsized events. In the past, a local disease outbreak was more likely to remain localised. Today a local disease outbreak can spread and cascade very quickly through various interconnected and interacting systems. However, it is also a feature that complex systems are self-organising and adapt. Instead of flying to another continent to meet a client, people video-conference. Instead of importing intermediate inputs and components from a abroad, they begin producing it locally. The system adjusts, albeit at a cost in terms of efficiency, but it adjusts nonetheless. If it does not, it is not a self-organising, adaptive system (Miller & Page 2007).
 |
| Commercial Air Traffic, 2009 |
Undoubtedly, the increased interconnectedness of the contemporary globalised world has created significant risks and vulnerabilities as well as opportunities (Keohane & Nye 1977, Bostrom 2008). Strong interconnectedness creates vulnerabilities, especially if there is a lack of redundancy build into the system. Complex systems theory suggests that diversification can help create greater resilience. If global chip production depends on only one or two countries and moreover on just-in-time supply management, the risk of catastrophic failure increases, especially if no "back-up system" is put in place. Moreover, the interconnectedness of the various sub-systems (production, trade, logistics, health, demand, economy, financial markets and politics) means that small events can cascade throughout other sub-systems making such events potentially even more impactful. This added level of “complexity” makes it even more difficult to predict the behaviour of the system or its sub-systems. It is possible to model the spread of a virus under certain assumptions and assign probabilities to various scenarios. But this is of only limited usefulness when it comes to predicting the reaction of economic agents or financial investors.
The impact of an epidemic also critically depends on agents' perceptions and especially their perception of risk. The coronavirus epidemic may be example of what Gerd Gigerenzer calls a ‘dread risk’, that is, a low-probability and high-impact event (at the individual level). Dread risk leads people to make mistakes. More people die from lightning than from terrorism, yet people take greater precautions vis-a-vis the risk of terrorism than lightning. Dread risk also related to the well-established availability and recency biases. Because more people avoided flying following 9/11, car accident related deaths increased and outstripped the total number of passengers killed on 9/11 (Gigerenzer 2004). Admittedly, given the initial uncertainty about the potential severity of risk, it may be prudent to over-estimate risk and resort to extreme measures (e.g. quarantines) and accept adverse consequences in other sub-systems (e.g. economic activity). However, once a clearer risk picture emerges, more rational behaviour should emerge at both the individual and government levels. There is an additional problem. Bureaucracies are by nature very risk-averse (Downs 1964) and similarly politicians will tend to err on the side of caution, especially in view of media coverage that favours human interest stories over probabilities and trade-offs. (The media targets the very biases in news consumers that are detrimental to dealing with a situation in a rational way.) Cognitively, availability heuristics lead media consumers to overestimate risks.
Bureaucrats and politicians are bound to act in a risk-averse, loss-avoiding manner - even more so in times of uncertainty (another well-established biases established by prospect theory). This complicates a fully rational, purely economic cost-benefit analysis. The complexities and uncertainty involved in imposing various measures and their unintended, indirect and longer-term consequences further complicate decision-making. Framing is often crucial (Tversky & Kahneman 1986). A purely rational cost-benefit analysis would need to (1) quantify the risks related to the epidemic, including the (economic) costs of fatalities and (2) quantify the costs of various counter-measures. Valuing a human life in monetary terms is emotionally uncomfortable. However, it can be done. Individuals could put a monetary value on their own life (bets, expected value) or insurance companies can calculate the monetary value of an individual in terms of discounted future cash flows. It is an emotionally uncomfortable but relevant question to ask: if it were to be established that the present epidemic is 10x more lethal than seasonal influenza, does this justify taking severe measures to combat the epidemic (BBC March 2, 2020)? And if a 10x higher mortality justifies the measures, would next year's seasonal influenza that is 2x as lethal justify taking the same measures again in spite of immediate economic costs and unintended consequences, including fatalities (e.g. people in quarantine not being able to access critical medication)? This a moral and ethical as well as economic question. It should be part and parcel of a rational public debate and government decision-making.
Historically, catastrophic disease outbreaks have occasionally had massive macro consequences, in addition to the loss of human life. Again the complex systems framework is useful. A small mutation in a relatively harmless virus can have massive, large-scale effects. The Spanish flu, for example, is estimated have killed 10-100 million people in 1918-20. Black death during the middle ages led to a sharp population decline and subsequently led to a rise in labour income and potentially helped reshape the political systems of the early modern era. Malaria may have been an important factor contributing to the fall of the Roman empire (BBC 2017; Harper 2017). Guns, Germs and Steel (Diamond 1999) explains why Western Europe an expansionism was so rapid and how disease agents helped it defeat large empires (see also McMichael 2017). European colonialism killed more indigenous people through diseases than through guns or steel. some studies suggest that disease agents imported by Europeans may have killed 90% of the pre-Columbian population in the Americas or 10% of the world population. It also contributed to global cooling by way of carbon absorption by arable, but uncultivated land following the demographic collapse. This in turn was a factor contributing to subsequent famines and rebellions in Europe and Asia (Koch et al. 2019). By comparison, WWII is estimated to have killed 3% of the global population. (What happened to the Incas might well be what happens to humans once extraterrestrial lifeforms do show up on planet earth.)
The existence of disease agents also helps explain long-term socio-economic processes and outcomes such as economic development. Economists have debated to what extent institutions, culture, natural resource endowment, geography or disease agents have affected countries’ economic development. The existence of certain diseases (e.g. malaria) may have had several detrimental effects. It may have negatively affected local health conditions and the indigenous population. It may also have prevented settler colonialism and favoured extractive forms of colonialism, which, in turn, influenced long-term institutional and economic development through presence/ absence of institutions conducive to economic development (Acemoglu, Johnson & Robinson 2000, Acemoglu, Johnson & Robinson 2001, MacArthurs & Sachs 2001, Sachs 2003). Think New England vs South Carolina, or Bahia vs Rio Grande do Sul, or New Zealand vs DRC). One does not need to take side in this debate in order to appreciate that disease agents can have significant macro-, even world-historical effects even if they, like other explanatory factors, are rarely the only causal factor affecting such outcomes.
The present epidemic is bound to trigger renewed intellectual efforts to understand and deal with future epidemics. Epidemologists and economists should think harder about how to design a close-to-optimal policy responses. How effective are quarantines compared to more voluntary measures to contain the spread of diseases? What are the likely costs and benefits of various measures? Should economic policy be geared towards monetary ro fiscal policy? Should fiscal policy be targeted (and targeting what) or broad? What measures can best help relieve supply chain bottlenecks? Will introducing greater automatic stabilisers help soften the economic downturn? Should or rather will politicians push for broader healthcare coverage in order to reduce the risk of uninsured people by not being able to get medical attention form becoming diseases vectors? How can governments cooperate more effectively at the international level? How can corporates deal more effectively with the risks emanating from inter-connectivity by putting in place more resilient systems and response protocols? Lots of intellectual work to be done.
 |
| Global Trade Flows |
Last but not least, epidemics cannot be wished away and will continue to occur again and again. While epidemiology is very useful to evaluate the severity of an epidemic (Kucharski 2020) and suggest ways how to deal with it, the relationship between humans and disease agents is a biological arms. In spite of humanity's advanced understanding of micro-biology, the risk of epidemics will persist given the unpredictability of mutations and the lead time that is needed to develop vaccines. And bacteria (and viruses) will not dispels given human's symbiotic relationship with them. Estimates (recently revised) suggest that the human body contains about as many bacteria as human cells (Sender et al. 2016). Bacteria even shape human behaviour in important ways (Enders 2015). In short, once one starts to think of viruses and bacteria, one finds and sees them everywhere. Anthropocentrism is in the end just one particular and not necessarily a heuristically very insightful way of looking at the world - including the social, economic and political world.