The Bear’s Lair: Why wait 30 years for robot CEOs?

Jack Ma, Chief Executive Officer of the Chinese Internet giant Alibaba, opined last week that in 30 years even CEO jobs would be outsourced to robots. His observation caused qualms in the overstuffed hearts of the cognitive elite, who had thought themselves immune to such vulgarities as replacement through automation. My question is: given the abysmal quality of current corporate top management, what in the world would make us wait so long for that transition to occur?

As far as I am aware, Ma’s outburst is the first that has seen robots as a threat to the jobs of the corporate elite. But of course, most of their jobs are as outsourceable as everybody else’s. The incentive is greater too; the direct saving in salary costs, expenses and stock options costs from replacing a human CEO is in the tens of millions of dollars, far more than the saving from making some poor schmuck on the assembly line redundant.

True entrepreneurs have jobs that are not robotizable. They develop new products, and expand their businesses into new areas through new and previously unimagined technologies. They are also expert at getting a great deal of result from very few resources, and at reacting to entirely new and unexpected shocks. One cannot imagine machines/robots becoming good at these tasks in any reasonable space of time; it is likely that even in 50 years there will be areas of business decision-making that are entirely unamenable to robot penetration.

Almost all CEOs, however are not true entrepreneurs, though they aspire to be paid like them. They operate within relatively stable environments, in some cases highly competitive, in other cases much less so through sweetheart deals with governments. Their competitive situations do not change all that rapidly; for example the threat from Internet shopping that is now hollowing out many bricks-and-mortar retailers has been developing for two decades, since the foundation of Amazon in 1994. Human CEO’s thus generally have an extraordinarily well-paid job that is largely routine.

Human CEOs have amassed an especially dire track record in the last two decades. Whereas their compensation has soared far faster than overall U.S. output, productivity growth in U.S. businesses has notably lagged, indicating their failure to invest optimally. By and large, CEOs have been fooled by the “funny money”: policies of the Fed and other central banks, and have over-leveraged their companies, bought back their stocks, invested in pointless acquisitions, bought too much real estate and generally engaged in policies that benefit their stock options’ values far more than the values received by ordinary shareholders.

Most of the frantic round of mergers and acquisitions that for many CEOs constitutes activity are either obvious or futile. Mergers to cut costs in an area of declining activity are the former, as are “bolt-on” mergers to increase a company’s penetration in an attractive market. On the other hand, most mergers by which a company expands into an entirely new area are futile (because the company has no special expertise in the new area, and so achieves sub-market returns there) as are mergers designed to bulk up the company’s size and thereby increase the compensation of the CEO. Robot CEOs would readily be able to recognizer the obvious mergers and reject the futile ones; the process of analysis required is only moderately complex.

This brings out an important advantage of robot CEOs; their lack of conflicts of interest. By law, their programming would be forced to reflect the interests of shareholders as first priority, and to forbid them from breaking all the laws to which a modern corporation is subject. They would not be able to shade their observance of laws in the belief that “everybody does it” or that they would never get caught. More important, they would not be able to devote their corporation’s resources to their own pet projects, environmental or otherwise, nor would they be able to follow policies such as witless corporate aggrandizement, that rewarded the CEO rather than the corporation’s shareholders in general.

Robots, not being human, would not be paid; they would simply cost what they cost. Their programming, which would be determined by corporate law first principles, would give them guidelines for resolving all the contentious points that cross a CEO’s desk. They would not form close emotional bonds with people, and would hence arbitrate disputes between subordinates with monotonous fairness. Production, finance and control would be natural to them, and they would be little worse at HR than the average human CEO, with his tendency to play favorites.

Robot CEOs’ weakest point would probably be marketing, where they would have little clue what would appeal to the average consumer. However, this could be solved by a good human chief marketing officer, and would in any case be offset by robot CEOs’ ability in industrial marketing, where they would often be selling to other robots.

Robot CEOs would provide shareholders with better results than human CEOs for 95% of companies, 95% of the time. Their decisions would be shareholder-optimized, so in most cases shareholders would find themselves benefiting from robotization by far more than the savings in salary/benefits/options of a robot compared with a human CEO. Like self-driving cars, robot CEOs would never get drunk, criminal, greedy or prejudiced and so would do a better job than human CEO’s in almost all cases. Also, as in self-driving cars, in extreme cases they would fail catastrophically, but those cases would be fewer in number than for a human CEO. The lower level of failures would thus cause a further substantial overall increase in shareholder value, compared to a human CEO-managed company.

The robot CEO would work a 168-hour week or close to it. Like human CEOs, it would waste a lot of time in meetings and conference calls, but it would have additional reserves of time available that a human CEO would be lack because of his physical limitations, need for a social life and family, etc.

It would be possible to mitigate further the risks of a robot CEO by using a human non-executive Chairman. The Chairman’s job would be to spot opportunities and risks for the company that were not obvious, and to identify them sufficiently forcefully that the robot CEO took them into its calculations. By this means, most of the “self-driving car runs into a tree” events would be avoided, and companies that needed to make a sharp strategic turn would be able to do so. The non-executive chairman’s workload would be limited, perhaps to two days a week.

The ideal non-executive chairman would be older, perhaps retired from his full-time job, but with experience of a wide range of industries, not just that served by the company. The 1960s British ideal of a merchant banker chairman would work very well if the CEO was a robot (not that merchant bankers exist any more, but a similar generalist with excellent high-level contacts and a preference for good lunches over hard work would be very suitable.). Conversely, a retired human CEO as Chairman of the same company would be too limited in breadth of experience, and far too governed by the assumptions of the company’s industry, many of which would have become outdated by the switch to a robot CEO.

The technology is not quite ready yet for a robot CEO to be appointed to head a major public company. However, the cost advantage of robot CEOs is obvious, and with each unit potentially commanding an initial sale price of several million dollars, comparable to a human CEO’s salary and benefits, the rewards for the company that develops such a prototype would potentially be considerable.

The early adopters would probably be those companies with relatively simple operations that have suffered from human CEO ineptitude in recent years; the serial acquirer Mondelez (NASDAQ:MDLZ) might be one such candidate. In many cases, once a robot CEO has been designed and bench tested, shareholders of companies that have suffered from inept management should feel that it could not do any worse than the humans under which their wealth has been eroded and opportunities missed, and would very likely do better.

I have written previously that our economy, if not afflicted by excess regulation, should be able to tolerate the rise of the robots without suffering mass unemployment, just as it did earlier technological advances. However, among the jobs most susceptible to robot replacement is that of Chief Executive Officer. The benefits of robotization to shareholders are just too great. I confidently predict that we will see the first examples well before 2047 and that in general they will be highly successful.

(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)