In our contemporary world of hyper-uncertainty, where we are being constantly surprised (and often upset) by unexpected outcomes, data would appear to be our friend. The more information we collect and interpret, the better we can analyse the past and the more certain we can be of the future. Data reduces uncertainty.
Not.
The fallacy of data is that far from reducing uncertainty, it is a significant cause of hyper-uncertainty. For example, opinion polls are vast collections of data, carefully selected to be representative, which are supposed to predict future events with more certainty than guesswork or gut feel. But currently we are exposed to all the problems which result from hyper-uncertainty precisely because the polls have proved to be wrong over and over again. In the 2015 UK General election. In the October 2016 Colombian Referendum to approve the peace deal with Farc. In the 2016 UK Referendum on membership of the EU. In the 2016 US Presidential Election. And quite possibly this weekend in the Italian Constitutional Referendum.
Of course, there is no news in the old truism that Rubbish In = Rubbish Out; that if you collect the wrong data, or you misanalyse it, the resultant knowledge will be useless. Except that opinion poll data is managed and interpreted by experts who are supposed to know exactly what to collect and how to analyse, but who are getting it wrong over and over again. And it is this failure of data to be useful which causes hyper-uncertainty; just how scary is the world if you cannot trust the data on which we base our decisions?
Yesterday the UK Government published a consultation document on executive remuneration, the central element of which is the that companies would be forced to publish the gap in pay between their CEO and the average employee. The result of this should be that shareholders and employees will be able to compare the executive pay gap across lots of different companies and therefore prevent excessive anomalies in their own organisation. In other words, the theory is that the collection and interpretation of this data will lead to an improvement in the ethical and equitable relationship between the fat cats and the hoi polloi.
But in the era of hyper-uncertainty this is unlikely to be the case. Because we won’t trust the data. For example, what do we mean by Executive Pay; does this include stock options, benefits in kind, bonuses? Do average earnings include those of workers who are outsourced, or part time, or zero hour contracted? In addition, faced with the regulatory requirement to publish this information will companies seek ways to cloud the transparency which it seeks to bring by finding new and clever ways of changing the rules.
So in practical terms the increase in data will actually lead to a further erosion in the trust between employers and employees and the Trades Unions who represent them. In an environment where data cannot be relied upon emotional responses becomes more powerful. An example of this is the dispute between Southern Rail and the ASLEF union over driver-only operated trains. The union claim that this is a safety issue, but data collected over the last 20 years indicates that driver-only operated trains, common in much of the rest of the UK rail network, are no less safe than trains which also have a guard. Of course part of this dispute is a power play between the parties (and the government who are pulling the strings of the employer). But one kernel of it is the ability of the union to exploit an uncertainty in the travelling public that driver-only trains might be less safe, no matter what the data suggests, simply because increasingly we distrust the data which our experts tell us is true.
At a tactical level negotiators have always needed to question data provided by their negotiating partners and identify issues and errors. But now, it might be enough just to negate the value of any data simply on the premise that all data is suspect. A tactic which might be great fun for some, and very frustrating for others.
Stephen White