Interestingly plausible but unproven

From How Corporations are Managed by Arnold Kling.

I have pointed out many times that there is a strikingly common zone of gender representation in all sorts of unrelated fields here in the US. Women make up 15-30% of the top achievers in just about every one of those fields if you measure achievement by awards, or compensation, or hierarchical position. Even in fields which are overwhelmingly female, the top performers skew male.

Many first order thinkers see this and assume that it is a product of simple bias and discrimination. But that explanation has come under assault as it has become increasingly clear that in the two areas of greatest possible influence, hiring and compensation, women are treated the same as men. The most deeply researched question is about unfair compensation practices. It is pretty compellingly demonstrated, and with frequent replication among different and independent research teams, that women are paid the same for the same work. Average compensation differences arise not from bias and discrimination but from different individual choices about fields of endeavor, appetite for risk, hours worked, continuity of careers, commitment, etc.

If men and women are being treated fairly in hiring and in compensation, it would be somewhat surprising to find some other less obvious mechanism of bias which winnowed them from top achievement.

Virtually all the public policy recommendations to address this perceived but not demonstrated inequity are pro-natalist oriented. Childcare, family leave, employment guaranties, etc. Ironically, those countries where these policies are most commonly and most extensively implemented also have the fewest women top achievers in far fewer fields of endeavor. The US is a benchmark in generating women high-achievers in many fields.

I am comfortable with this overall explanation - that men and women who are willing to take considered risks, work long hours, consistently over several years and with great commitment (working extended irregular hours) are the ones who rise to the top and that the differentials at the top reflect different willingness to adhere to those requirements.

I have not overly focused on what causes the breadth of the range of 15-30%, assuming that it reflected different degrees of demand in different fields.

Kling has a different and plausible speculation. I don't think we have the data to support it. Yet. Indeed, I think, plausible as the hypothesis is, that the hypothesis is the product of a series of assumptions, a couple of which are plausible but unproven.

He starts out by making the point that differences in enterprise scale drive differences in executive function.

Think of a simple restaurant, with about a dozen employees. The owner-manager knows every employee and how they are doing. The owner-manager can observe how customers are experiencing the food. The owner-manager can be on top of every aspect of the restaurant, from caring for the facility to setting the menu to managing relationships with suppliers to food preparation and serving. In short, the owner-manager can make decisions based on what is directly visible.

In a large corporation, the CEO has never met most of the employees. The CEO observes very few customer transactions. The CEO is ignorant of many of the functions that support the business. Compared with the restaurant owner/manager, the CEO has a tiny fraction of the information pertinent to running the business. What information the CEO does obtain comes mostly indirectly via memoranda, briefings, and management reports.

I am good with that observation.

The next link is experientially true but I am unfamiliar with how well it is measured and demonstrated. We all know successful people who cannot move beyond sequential linear thought processes of IF-THEN. We also know people who can produce useful insight through processes of deep abstraction, who see that which most do not.

Some psychologists draw a distinction between concrete and abstract thinking. Concrete thinking means that your mind works only with the objects that you see and feel around you. Abstract thinking means incorporating imagination and theorizing into your thought process.

Because the owner-manager of the restaurant can see everything that is relevant to business decisions, the owner-manager can operate mostly on the basis of concrete thinking. The owner-manager does not have to constantly update and revise a “theory” of the restaurant.

But the corporate CEO, operating with a limited information set that arrives indirectly, must use more abstract thinking. We may think of the CEO as trying to navigate in a confusing forest using only little scraps of a map. The CEO operates with a theory of the business and fits those little map scraps into the theory.

I am struck by this observation; that there are different degrees of legibility at different scales of business. It is obviously true. It is obviously consequential. And yet it is little discussed.

He describes the consequentiality thus:

So far, I have suggested that corporate CEOs know less about their organizations than small business owner/managers do. Does this mean that corporations should always lose out to small businesses?

In fact, of course, large corporations have major offsetting advantages. An important advantage is that there is a lot of cultural capital embedded in a corporation. As a corporation evolves, its employees and managers acquire habits, customs, procedures, relationships, and techniques that are valuable assets.

When the business environment rewards accumulated cultural capital, expect large corporations to do relatively well. But when the situation makes it paramount that leaders make well-informed decisions, expect a lot of bad management and failure, because good information is scarce within a large organization. It is really difficult for corporate CEOs to avoid messing up.

Again, comports with my life and professional experience.

The next link in his argument is probably the weakest, not because it is untrue but because it is insufficiently proven.

The “glass ceiling” describes the phenomenon that relatively few women reach the top position in large corporations. Most people assume that this is due to discrimination.

As a possible alternative explanation, suppose that the very high level of abstract theorizing that is required of CEOs, while found rarely at all, is relatively more prevalent among males than among females. This is related to the controversial theories of male-female average brain differences suggested by Simon Baron-Cohen. As explained in Wikipedia,

In 1997 Baron-Cohen developed the empathising–systemising theory. His theory is that a cognitive profile with a systemising drive that is stronger than empathising is associated with maths, science and technology skills, and exists in families with autism spectrum disorders. . .

Baron-Cohen’s work in systemising-empathising led him to investigate whether higher levels of fetal testosterone explain the increased prevalence of autism spectrum disorders among males;[18] his theory is known as the “extreme male brain” theory of autism.[12] A review of his book The Essential Difference published in Nature in 2003 summarises his proposal as: “the male brain is programmed to systemize and the female brain to empathize … Asperger’s syndrome represents the extreme male brain”.[19] Critics say that because his work has focused on higher-functioning individuals with autism spectrum disorders, it requires independent replication with broader samples.[20] His prediction that prenatal testosterone would be elevated in autism has been confirmed.[21]

What Baron-Cohen calls systemising may correspond to a propensity for abstract thinking in business. In that case, that could explain the predominance of males as CEOs.

Kling is suitably cautious with his own hypothesis.

I am not at all confident that this difference in thinking styles truly explains the “glass ceiling.” I only offer it as an intriguing possibility, not at all proven.

There is useful discussion in the comments section.

I think he is right to be cautious. We know most of the differential is due to work requirements intersecting with willingness and capacity to considered risks, work long hours, consistently over several years and with great commitment (working extended irregular hours). That explains the fact that there is a 15-30% differential.

But I wonder whether Kling hypothesis might explain differences within the 15-30%. Not all top achievements occur in a conext of large hierarchies. Artists, chefs, technicians, musicians, most services fields, are usually individuals within a narrow domains. Kling's hypothesis would have relatively less relevance there.

There tend to be two primary classes of large enterprises - commercial and government. They require a significant degree of abstract thinking as described by Kling.

But there is a further dimension for both these categories. Commercial and Government enterprises are more or less dynamic and uncertain. Government agencies tend to be far more stable and predictable than are commercial enterprises (with some rare exceptions.) Within commerce there are some sectors which are far more dynamic and unpredictable than others (think of the tech sector compared to the utility sector or the retail banking sector.)

My anecdotal impression is that we do see high-achieving women more at the 30% level for staid public sector agencies and for more stable commercial enterprises. It is also my impression that women are much less represented, more at the 15% level in the tech sector, investment banking, and other such high uncertainty businesses where abstract thinking is at a premium.

Perhaps work requirements is what drives the 15-30% window but abstract thinking capability drives where a sector will fall within that 15-30%.

I am willing to accept the implications of the direct experience model of business versus the uncertainty condition of large enterprises and therefore the differential need for systemitizing, abstract thinking, imagination.

Where I struggle is with Baron-Cohen's hypothesis that there is a differential in abstract thinking between women and men. And to be clear, the issue is not whether or not women can think abstractly. Of course they can and do. The issue is whether there is a material difference in inclination towards systemitizing thinking between women and men, with men being more inclined towards systemitizing.

I can think of all sorts of personal experiences and fields of endeavor which support that presumption. But I am unaware of any sort of systematic study (yes, I know) which documents that perception.

It is an interesting hypotheses on Kling's part. TBD