Lessons in resilience from daily life under COVID-19

“At least we haven’t lost internet coverage”

Working remotely from my home in the San Francisco suburbs has been a much smoother transition than I had imagined, thanks in no small part to the availability and reliability of internet bandwidth, even in the face of huge increases in the use of data intensive video applications. This stands in stark contrast to the extreme stresses and gyrations I observe around me in the healthcare system, the education system, the electrical supply system and in other supply chains.  More than once friends and colleagues have remarked on our good fortune to have access to reliable internet. 

Of course, this is no mere accident. The internet was originally constructed for military use and was explicitly designed to withstand a nuclear attack. It does so by having a modular decentralized design so that attack on any part of it does not trigger the failure of the whole. The network has massive structural redundancy, such that messages can be automatically rerouted according to data traffic conditions. And error correction (another form of redundancy) assures functionality even if there are transmission errors. It is the very “inefficiency” of the design that has enabled us to smoothly and rapidly adapt to new and unforeseen circumstances. The architects of the internet employed prudence by designing for a wide range of future conditions, rather than for short run efficiency. 

In stark contrast, electrical power generation is highly centralized, driven by massive economies of scale. And while transmission grids were adopted in the early part of the last century to enable load spreading and to encourage competition among generators, the radial design is mainly driven efficiency considerations rather than resilience. Because of this, black outs and brown outs are not uncommon. Across California we have recently experienced rolling “load shedding” outages due to the inability of the system to adapt to the higher demand levels triggered by a heatwave. We are effectively paying for lack of past investment in buffering capacity. We can think of resilience as ensuring long run efficiency at the cost of short run efficiency – a trade-off between different timeframes.  

We are also expecting further electrical shutdowns for an entirely different reason: the risk of fires being sparked by powerlines under the increasingly hot, dry and windy climate of the region, a factor which would have been hard for the system’s original architects to foresee. Resilience is always relative to the specific risks encountered and these change over time. Redundancy, prudence and modularity must therefore be complemented by adaptability – the ability of a system to evolve and learn over time. 

Every system, even the California power grid, must eventually adapt to altered circumstances if it is to survive. As I look out across my neighborhood, I observe the slow spread of solar panels, storage batteries and back-up generators in response to unreliable power supply. Over time this adaptability in embracing a greater diversity of power sources, increased redundancy of generating capacity, and the creation of a more decentralized modular structure will hopefully bolster reliability. Resilience also therefore concerns to the speed with which functionality is restored after a disturbance. Considering the markedly differential recovery rates in the performance of different companies, industries, states and countries, we are effectively all witnesses and participants in a huge experiment on the resilience value of different policies and strategies. The lessons should be plentiful should we care to observe and register them. 

I read in this morning’s news of airlines planning to shed thousands of jobs in response to depressed demand: a painful but necessary adaptation for airlines, with their narrowly specialized business models. While the business model of airlines may be quite brittle in this respect, aviation has in other ways shown remarkable resilience. Over the course of a century it has gone from being one of the riskiest forms of transport to the one of the safest. This is in large part due to the accident investigation system, whereby every accident is investigated according to a strict protocol, the results made public and mandatory recommendations made and adopted, ensuring the adaptability and improvement of the system over time. One can only speculate about the enormous impact such an approach could have if applied in other walks of life – from medicine to power distribution. 

In the Bay Area, the ultimate resilience will however come with the next major earthquake. My house insurance excludes earthquake coverage, which can be purchased separately for a very substantial premium. Many choose not to purchase earthquake coverage or to insure only a proportion of the value of their properties. Resilience is a lot like earthquake insurance – it provides protection against future uncertainty, but this comes at a cost in the present, and begs the thorny question of how to determine the right level of “coverage”. Even if a precise value cannot be assessed, presumably coverage should at least be sufficient to ensure survival against plausible risk scenarios. 

Eventually things will normalize, and even if they don’t return exactly to the pre-crisis status quo, we are in danger of losing our current heightened sense of the value and importance of resilience. It’s worth asking ourselves before the crisis fades into a story which we will nostalgically tell to our grand children, whether we have managed ourselves, our businesses and our institutions for efficiency or for resilience, and whether and how we want to recalibrate that trade off given what we have learned. In this very real sense, the crisis is an opportunity to better protect ourselves against future ones.