Volkswagen is not alone or even the worst. Wittingly or not, almost all companies are built on lies and and self-deception, from top to bottom. Cheating is endemic.
Whether by financial or (in the VW case) physical engineering, executives routinely manipulate or ‘smoothe’ corporate results to meet stock-market earning expectations and trigger their own cash or stock-option bonuses. Recent research suggests that selection panels recruit top accountants who are willing to do just that over those who aren’t.
‘If you aren’t lying you aren’t trying!’ exhorted a Barclays manager quoted in the LIBOR-fixing enquiry. Lower down the organisation, people equally routinely cheat in small or large ways to make sales or other targets and keep managers off their backs. But small-scale cheating can all too easily morph into institutional self-deception and eventual scandal (the banks, Mid-Staffs, VW) in which, regular as clockwork, leaders protest their innocence and blame ‘a few bad apples’ for the misdeeds. At the extreme, as I’ve noted before, self-deception and denial can turn whole organisations into reversed-out versions of themselves: banks that make people poor, hospitals that make them medically dependent, shareholder-value-maximising companies that destroy value.
In view of the dire effects, why are lying and self-deception so prevalent? It’s not that companies are full of wicked people. It’s the system, stupid – as the telling phrase ‘gaming the system’ clearly announces. And the culprit? In a word, measures and the counterproductive ways in which managers use them.
Measuring is one of the most important things managers do. Without appropriate measures, management can’t be even methodical, let alone scientific. Yet given the lip-service paid them (how many times have you heard the expression, ‘What gets measured gets managed?), astonishing quantities of the measures that companies compile are useless or worse. While good measures promote learning, improvement and innovation, bad measures are often worse than no measures at all: not only do they drive actions that alienate customers and their own employees (yes, you do get what you measure), they keep managers in the dark about their effects.
The reason is simple and profound – but not obvious. As John Seddon has put it, whether organisations realise it or not, ‘there is a systemic relationship between purpose (what we are here to do), measures (how we know how we are doing) and method (how we do it)’. Measures should be a means to an end (learning, improving). When purpose is put first – doing what matters to the customer through products or service – and measures are related to it, managers and employees can see directly how well they are achieving that purpose. Their job is then, guided by the measures, to identify methods to do it more effectively. That is, the purpose-measures-method trio form a learning system. The job of measures is to connect actions with consequences.
Now look what happens when the purpose is not related to customers, or there is no explicit purpose at all. To fill the vacuum, the measure becomes the de facto purpose – the means turns into the end. But putting the measure first (as in sales or other targets, standards and specifications mandated by governments and regulators) has terrible consequences, because methods are now geared to achieving the measure rather than the real purpose (often at the expense of the real purpose). They are not about learning but making people accountable. And as all such targets are arbitrary, with no necessary relation to the system’s capability to meet them, there are high incentives for people to cheat. As W.E. Deming succinctly put it half a century ago: ‘People with targets and jobs dependent upon meeting them will probably meet the targets — even if they have to destroy the enterprise to do it’.
Cue Volkswagen, seemingly a classic case of capture by its own closed, self-referencing metrics. Fifty per cent owned by the Porsche family and 20 per cent by the state of Saxony, VW, as the company at the centre the company town of Wolfsburg, was focused entirely on volume and jobs, to which environmental protection and even the customer came a distant third. Everyone seems to have bought into those goals, including the supervisory board, the unions and shareholders who were doing very well from the formula. But volume meant selling in the US, which has much stricter diesel emission rules than Europe. With feeding the machine the priority, and detached from customers, it’s easy to see how self-confident, production-minded managers could convince themselves that cutting a few testing corners wasn’t a serious matter – after all, fudged emission tests don’t make them poor cars, just not quite what the ads said they were (ads also lie, as consumers well know; witness the soaring use of online adblockers). Little lies become bigger ones, until it is very hard to go back.
As set out in Goodhart’s Law – ‘Any statistical regularity will tend to collapse once pressure is placed upon it for control purposes’; or in plain words, a measure that becomes a target is no longer reliable as a measure – all data collected to meet incentives is worthless because liable to manipulation. The mother of all numerical targets, of course, is shareholder value, which is useless as a guide to action in the real world of customers and service, because it is unrelated to what matters to customers, is used as a carrot or stick rather than an aid to learning, and does not connect actions to consequences. It is owned by the boardroom, not those doing the work. Incidentally, it may well be true that departed CEO Martin Winterkorn did not have direct knowledge of the testing scam: in a delicious variant on Catch-22, in an organisation that uses measures to make people accountable rather than to learn, no one is ever accountable because they always meet their targets. Winterkorn's deeper responsibility, of course, is for the system that drove such perverse behaviours in the first place.
As VW’s subsequent struggles show, the distorting effects of turning measures into ends are catastrophic. The company has set aside $6n to provide for vehicle recalls and rectifications, while liabilities for fines and possible lawsuits could dwarf that. Some executives may face criminal charges. If, back in the real world, car-buyers give up on the brand, things could get even worse. Spending is already being reined back. All of which shows that actions do eventually have consequences, even if the failure of official measures to make the link apparent means that they turn up many years later as an unexpected and very nasty surprise. Rarely have the wings of chickens flapping home to roost cast a longer or deeper shadow.