Saturday, March 15, 2014

The Irrationality of Corporate Decision-Making

Corporate decision-making is hard to explain. It is not rational, not even approximately rational. Many people make the mistake of thinking that corporations are rational in financial matters, but fail to balance financial priorities with other priorities. But it is easy to find examples to show that even when only financial consequences are considered, decisions are still not rational.

Two of the top news stories this week are about the consequences of weak corporate decision-making. One story is about Target, one of the largest U.S. retailers. It ignored initial warnings of spyware on its networks, reacting only after records of 80 million customers had been copied out to unknown destinations. The other story is about the investigation into decisions that led to a design defect being incorporated into more than 1 million cars manufactured by General Motors.

We don’t know the details of the thinking at Target, but it sounds as if less than 1 hour, or less than $100, was spent looking into reports of the network intrusion. At that point, it seems the intrusion was noted and logged, but there were no specific plans to investigate further or take actions to correct it. Target seems almost certain to spend more than $1 billion in technology costs alone in its belated response to the problem. So even if the initial reports indicated only a slight chance of a major problem, the scale of the initial reaction was not in any rational proportion to what was at risk.

At General Motors, the design defect at issue was noted and discussed, and a solution was proposed that would have cost an estimated $1 per vehicle. This was rejected at the time as being too expensive, so instead, the defect was built into three years of cars. Investigators both inside and outside the company are trying to find out how that decision could have been made. The cost of the preemptive fix would have been tiny compared to what General Motors now must pay for a recall, which is on the order of $200 per vehicle. Looking at the decision in the narrowest financial terms, it shows a disregard for consequences that cannot be described as rational.

Corporate decisions in general reflect a resistance to change. The actions of change are regarded as more risky than they are, but it goes deeper than this. Information that change is needed or that change is taking place is also resisted. Target apparently did not want to take even one cash register offline to find out the nature of the spyware that had been installed, presumably because interrupting a cash register, even one known to be corrupted, was considered too big a change to make. General Motors perhaps did not want to reopen a design that had already gone through the painful process of approval at approximately ten levels of management. And these are not bizarre, extreme cases of corporate decision-making, even if the negative consequences happen to be unusually large and extremely visible. Large organizations every day make decisions that put their own financial future at risk in order to avoid thinking about change. The risk of financial failure that corporations take on is not merely theoretical, as you can see from the parade of large corporate bankruptcies.

The resistance to change is just one of the more obvious irrational properties of corporate decisions. The division of attention that goes into the corporate structure makes it impossible to keep considerations in their proper proportions. It is said that a corporation can approach rationality in the long run, but that is only a theory, and it depends on the assumption that the corporation is able to change faster than any changes that take place in the environment it operates in. All evidence and the very nature of the corporation point to the contrary. Once you recognize the built-in irrationality of corporate decisions, you can deal with corporations more constructively than if you imagine that corporations must be making decisions that reflect their own interests if those interests are defined narrowly enough.