leadership in complex adaptive systems

This is a reflection on Heifetz’ model of Adaptive Leadership, a practical prescription from HBS on becoming an adaptive leader in uncertain times

The Heifetz model of leadership adaptation requires an ability to develop a theory of action, rules of cause and effect that allow us to diagnose a situation and then select an appropriate action, suitably resourced, that incorporates others etc. in order to achieve a more predictable success.

The article you reference shows pretty clearly however that the skills traits and characteristics of leaders in a CAS are different than that of less complex organizations. The idea that simple behavior can create emergent phenomenon that are unpredictable defeats the idea of control systems and cause-and-effect. In the military, we are experiencing this phenomenon with respect to the complexity of nationbuilding in the land where the nationstate is not the natural organizational structure and yet it’s the one that we are trying to compel.

In many ways, culture seems to me to be one of those emergent phenomenon that takes on a life of its own in ways that nobody realized, understood or predicted.

It’s hard to reconcile the idea of independent agents with the usual models of leadership and even Heifetz is model of adaptive leadership since independent agents do not grant authority to others except on a transactional basis.

CAS with high degrees of volatility and uncertainty do not lend themselves to routine diagnosis since every situation may be a data point of one, never to be repeated. These kinds of systems can only be appreciated not understood, and certainly not controlled in the Frederick Taylor sense.

Re: Tamara’s question about decomposing the health care system and decomposing CAS more generally: the answer has everything to do with how tightly coupled the component subsystems are in the CAS.

Loosely coupled networks can allow for local subsystem management because there is a time delay between the changes you propose and the reflective effects from other subgroups affected by your change. The amount of interconnectedness, or the number of nodes that share a direct relationship with the subsystem under study, also influences how much the subsystem can be studied as a separate entity.

As an example, a house of cards can be said to be very tightly coupled and very interoperable with little to no slack. This is a structure that does not lend itself to subsystem management. A different example would be a set of dominoes in which you can create subsystems of dominoes so that if one falls over it doesn’t knock down the entire master system but only the subsystem that you have isolated by removing key connectors. This is how those giant domino structures are built and only at the end do they add the crucial connectors to create the master system.

Dr. Scott Page from the University of Michigan has done some excellent work in this area by identifying four different elements by which we can consider the complexity of a given system.

The elements of complexity that you described give caution to Heifetz is indicators of system underperformance which oversimplifies the challenge of anticipating future underperformance.

In fact we may be performing to the same level in the present while the environment changes around us and the changes have not yet manifested in our performance levels. If we wait until we have a series of underperformance as, it may be too late to begin the adaptive leadership that he prescribes. In general, I find Heifetz too reactive, which is an irony for a book on adaptive leadership.