Data and models - narrative takes them to the masses

I have been an advocate and promoter for the application of data science to marketing and communications for quite some time, but never really classified myself as a ‘numbers person’. There are many more inputs to success, including experience and creativity - dare i say it, even artistry. This is the human software brought to strategy and decisions that is impossible to document or replicate.

If the current health/economic crisis has done anything, it has reinforced the role of narrative in taking data science to the masses. Paradoxically, it appears that our political class is a convert to being guided by science on COVID strategy, but not on climate change. In both instances, the models will inevitably prove imperfect, but they’re proving to be close enough - particularly ‘for government work’ as the saying goes.

The paradox is created not by data, but by the narrative surrounding it. For all intents and purposes the messaging around COVID has been synchronised. I choose this term because consistency across tiers of government and their respective roles and responsibilities is too strong a word.

Data and interpretation by expert panels has been critical to the apparent success of COVID-19’s suppression within Australia. Collaboration across tiers of government through the laudable formation of the National Cabinet was the first step to achieving unity of purpose and message.

It enabled pooling of data and expert advice, access and transparency across jurisdictions, the evolution of a broadly agreed strategy and the content and timing of the messages explaining it. Again, let’s overlook some inconsistencies of narrative and event timing across state boundaries. Relative to what we’re seeing in other parts of the world, Australia is performing at Olympian standard on this.

There are many uncertainties about how we will emerge from the COVID-19 pandemic, with an amazing array of newly minted pundits talking its impacts on society, economics and our mobility around the globe. Unknown also is the future direction of the debate around climate change.

In Australia at least, climate and energy policies have far and away been the most toxic political issues. they have been the driver of more frequent leadership changes than ever previously experienced. Had it not been for COVID-19, it would undoubtedly have continued to take centre stage. Bizarrely, both sides of that debate have also used science and/or its validity to support their positions.

Hopefully, COVID-19 will end the debate about mathematical modelling associated with climate change. We are experiencing decision making based on epidemiological modelling during the COVID-19 crisis. The intersection of this data and economic modelling forms the fulcrum around which relative risks are considered and decided upon - the point at which we take a greater risk of virus spread in the interests of cranking up the international economy.

Notwithstanding this, the debate over the validity of climate change modelling will likely be rekindled as the cleaner environment resulting from the coronavirus shutdown returns to the hazier and more sluggish looking days of normal operations.

This despite the fact that in January this year, NASA reported that the modelling is proving correct. Other reports have suggested that the rate of change predicted may indeed be proving too conservative. Where the successful interventions into the COVID-19 crisis have been early and hard, the world has failed to act in a concerted and aggressive manner to reduce carbon emissions.* Take a look at Brazil and the USA if you want to understand what COVID might have looked like if the modelling had been treated with is as little urgency as the climate change models.

The problem with data and modelling is that it is based and is variable on a variety of assumptions, interventions and, significantly at times, unforeseen events.

One of the variables affecting modelling is agreement on the integrity of the model itself. The Imperial College modelling by Professor Neil Ferguson, which projected COVID-19 deaths in the UK at 510,000 as a ‘reasonable worst case’ has been widely discredited^, ostensibly by Ferguson himself when he later testified to the UK Parliament that the number was more likely to ‘top out’ at about 20,000.

But in an article in The Atlantic** on 2 April 2020, Associate Professor Zeynip Tufekci, argued that Ferguson’s apparent turnaround was not an admission of error, but was actually at the lower end of his initially modelled projections:

“[But] there was no turn, no walking back, not even a revision in the model. If you read the original paper, the model lays out a range of predictions—from tens of thousands to 500,000 dead—which all depend on how people react. That variety of potential outcomes coming from a single epidemiological model may seem extreme and even counterintuitive. But that’s an intrinsic part of how they operate, because epidemics are especially sensitive to initial inputs and timing, and because epidemics grow exponentially.”

The release of the initial report prompted the UK government to announce a far-reaching lockdown with significant economic consequences. If the lockdown had not occurred, what would the UK’s infection and death rates looked like? We will never know.

One of the reasons we will never know, according to Tufekci is:

“When an epidemiological model is believed and acted on, it can look like it was false. These models are not snapshots of the future. They always describe a range of possibilities—and those possibilities are highly sensitive to our actions.”

The difference in approach to the use and even the acceptance of data and evidence-based decision making between COVID and climate is most likely down to the former being a clear and present danger, one that can be most easily translated into a narrative from which almost every human on the planet can make decisions based on personal vulnerability and circumstance.

This is no creeping outcome. You could catch the novel coronavirus tomorrow and be ill within days. The impact of climate change is less discernible, unless you’re one of the many Australians dealing with bushfire recovery, the international communities dealing with more intense and frequent storms or droughts, or oceanic islanders seeing their homeland diminishing ever so surely.

Narrative is key. I have consumed two books recently. Radical Uncertainty’ is co-authored by two economists - former Bank of England Governor, Sir Mervyn King, and former Financial Times journalist, John Kay. The other is a belated read of Sapiens’ by historian, Yuval Noah Harari. You could not have more diverse authorship and perspective, but they both distill the direction of human history and destiny down to narrative as a fundamental driver.

King and Kay argue that data is limited in its scope to deal with radical uncertainty, events like COVID-19 that throw carefully constructed mathematical models and their assumptions out the window. Models are ‘small world’ views constrained by assumptions and constructs that, by definition, preclude ‘large world’ events.

“Our own experience in economics is that the most common explanation of a surprising result is that someone has made an error. In finance, economics and business, models never describe ‘the world as it really is’. Informed judgement will always be required in understanding and interpreting the output of a model and using it in any large-world situation.” ^^

It common language, what they’re really saying is what anyone involved in research and analytics knows - data is not insight, which is derived from intelligent interpretation of the data to create a narrative to explain the past or present and, if we’re really smart and lucky, articulate a path for the future.

Look at global investment markets in 2020. For months, leading investors had been warning of a looming debt crisis and that stock prices were at unsustainable and unjustifiable levels. The markets pushed upwards into February 2020 to record levels.

Unfortunately, the emerging narrative was not powerful enough for investors to exercise greater caution. They continued behaving as if there was no tomorrow, but eventually tomorrow came and with it an unwinding of the global economy as we knew it. Markets caved in, equity prices dropped more than 30%. The narrative and its tone had changed thanks to the undeniable economic impact of COVID-19 and become overwhelmingly convincing. Cash was king, at least for a month or so until a mix of potions like the US President confirming his regular fix of an unproven anti-viral treatment, the flattening of curves and the green shoots of a return to business were added to the story.

This has been an extraordinary time for great communicators to emerge and the veil to be pulled back on the shortcomings of others. To be fair, the unfolding pandemic has been a new experience to all involved and the wisdom of some of the actions taken will only be confirmed in the months and years ahead.

Members of the communities led by those who have completely ignored the data and specialist advice on the outbreak have suffered greatly, sometimes fatally. The best leaders have understood the data but, most importantly, have clearly articulated the risks, actions to be taken and the way forward. They have built a narrative that not everyone may agree with, but most can understand and follow.

it’s easy to be critical of those who fail to build a an effective narrative. Harari remarks on this in Sapiens:

“Telling stories is not easy. The difficulty lies no in telling the story, but in convincing everyone else to believe it. Much of history revolves around this question: how does one convince millions of people to believe particular stories about gods, or nations or limited liability companies? Yet, when it succeeds, if gives Sapiens immense power, because it enables millions of strangers to cooperate and work towards common goals.” ***

In the context of recent events, Greek philosopher, Plato’s attributed comment ‘Those who tell the stories rule society’ resonates loudly (noting that no one according to my research can actually locate the source of this quote and attribution to Plato).

* https://climate.nasa.gov/news/2943/study-confirms-climate-models-are-getting-future-warming-projections-right/

^ https://www.theaustralian.com.au/world/coronavirus-uk-modelling-that-forced-lockdown-messy-as-angel-hair-pasta/news-story/c7a5ea882a1a4066333d998e2a206f48

** https://www.theatlantic.com/technology/archive/2020/04/coronavirus-models-arent-supposed-be-right/609271/

^^ John Kay and Mervyn King, Radical Uncertainty, The Bridge Street Press, 2020. p261

*** Yuval Noah Harari, Sapiens, Vintage, 2011. p35