Domain Shift

Brian Arthur tells us that technology most often advances in the form of domain shifts.  In his narrative technologies cluster in related groups he calls domains.  So individual technologies might advance through a tweak here or there, but the economy advances through a shift from one cluster of technologies to another.

I like this idea.  Especially when we then broaden the topic to use technology as a background against which we view a particular slice of history.  Thus we can think, legitimately about the “steam era” and such.  Adding insight to this, Arthur situates all technologies as extensions of natural phenomena.  So he characterizes a technology as a phenomenon being “captured and put to use”.

All good.  This gives us a practical definition of technology and helps map its evolution onto the changes of society itself as it became more and more infused by technology.

Arthur also tries to make us embrace a very broad notion of technology.  He insists that a great number of social structures can also be defined as technologies.  So a modern business organization is a technology, as is any other institution.  I agree.  When we go further and identify that most, if not all, technologies can be segmented into sub-technologies we start to see commonalities that might elude us in a more holistic analysis.  This does not imply we can predict the higher level organization from the sum of the component parts, but it does allow us to see common ancestors to many of our contemporary technologies: they are rooted in lineages in the same way that the natural world is.

Let’s stay with this.

If a modern business corporation is an instance of a technology and is comprised of sub-technologies drawn from a common domain then we can identify the notion of shareholder value as one such component.  It is a technology.  That it has come to play such a prominent role in the overall construct of the corporation needs explanation.  How come other components didn’t reach the same dominance?

I was prompted tom ponder this by a short article by Daron Acemoglu in which he discusses the emergence and issues of what he calls our supply chain mess.

Extensive supply chains became the norm during the past few decades.  Corporations after corporation disaggregated it’s various processes and distributed them along these chains of activity.  Sometimes the chains were simply logical.  Most often they were expressed physically, and we ended up with a vast logistical network drawn across the globe each node of which depended on the operation of a multitude of others.  Just in time became the rallying cry of consultants and managers intent on squeezing what Acemoglu calls efficiency from the disaggregation of business activities.

Acemoglu sticks close to the standard explanation economists apply to this disaggregation.  By taking a business process apart and shifting some of it outside of the original firm’s boundaries so that it now occurs in a market context we are told efficiency rises.  This is taken as axiomatic.  Trade, in the standard telling, always increases efficiency.  The two sides both benefit in some way, so there is an aggregate gain.  Society is better off.

In the world of business management theory this breaking apart of business processes and the subsequent movement of pieces into the market is a consequence of the idea of core competencies first promulgated by Prahalad and Hamel back in the late 1980s and early 1990s.  I remember the emergence of this idea well: it caused quite a stir and was quickly taken up by all manner of firms intent on removing activities from their internal control that were not specific to their “mission”.

It is clear from this apparent coincidence between the standard ideas of economics as referred to by Acemoglu and of business as presented by Prahalad and Hamel that the notion of efficiency had jumped from economics into business practice.  Business theorists had adopted a key idea from economics and then turned into a practical measure to be applied by managers who wanted to be seen as cutting edge.  Our theories of the firm coalesced around this centrality of efficiency.  This is despite the controversial nature of what, exactly, we mean by efficiency.  More to the point: how do we even measure it in the face of endemic uncertainty?  Just what is an efficient firm?  How do we know when we hit that mark?

Of course, Prahalad and Hamel were riding a wave.  Their ideas were another indication of the invasion of business thinking by neoclassical theory.  Perhaps so-called agency theory is an even better instance.  It sprang to prominence during those same decades.  Its core idea was that there was an agent acting on behalf of a principal in many economic circumstances, and it then subjected that relationship to all the usual somewhat cynical  views of human relationships that populate standard theory.  Managers were suddenly agents of someone else rather than being administrators of an organization that was, itself, an independent being. And, given the neoclassical turn in economics at the time, they were agents of capital providers rather than anyone else.  Capital suddenly owned things rather than simply financing them.  Managers had to pay attention only to capital in the way described by Milton Friedman.

This brings us back to Arthur.  All these neoclassical ideas moving from economics into business practice represented a domain shift.  The previous set of technologies defining a business firm, its intentions, its social role, and the best methods for management, were all replaced by a new set drawn from the same root as neoclassical economics.  The underlying objectives of economics became the underlying objectives of business.  Instead of economics being the study of what goes on in the real world it became the source for what goes on.  It went from being a study to being an active agent in the landscape it sought to describe.  It became performative.  Economics, in this form, thus lost its independence.  It was designing the economy according to its current notions, no matter how controversial those notions might be.  Efficiency being one such.

So here we are in this tangled supply chain mess that Acemoglu describes but does not get to the heart of.  The mess is largely a consequence of the failure to account for uncertainty.  Which is an error of standard economics.  When, towards the end of his article, Acemoglu describes shareholder value as an ideology, he misses the opportunity to collapse standard economics into that same bucket.  After all the two occupy the same domain.  They co-exist as components within the same cluster of technologies.  If one is ideological, so is the other. The supply chain mess is a wonderful real world example of how economics fails in its relevancy to the problems that exist there.  The efficacy of global networks is subject to the uncertainties that all business faces.  The problem is that economics denies the importance of this uncertainty.  In its effort to simplify the world for tractable analysis it ignored real world features that turn out to be extremely important.  And when it then exported that simplification back into the real world in the form of business management theory it exported its naivety and led business astray.

Yes,  Arthur is correct.  Technologies are best thought of as clusters with common ancestors.  Shareholder value and neoclassical economics belong in the same family.  I just wish Acemoglu had been more clear in saying that.

One last thing. Talking of domain shifts: I wonder whether Arthur’s taxonomy needs a higher order or layer.  I agree that technological domains are a good analytical tool for studying the evolution of the economy.  But what of the larger scale?  What about a shift from a predominantly material economy towards a predominantly digital one?  If our institutions are technologies conceived to solve particular problems  and draw on a lineage of older technologies, what happens when a whole new lineage becomes possible?  What about the possibility of a technological Cambrian explosion?  One of the problems we face is that the opportunity to redraw our institutions based upon a strictly digital basis is being missed because we are over-invested in the older physical lineage.  The entire industrial era can be taken as one “epoch” manifested in all manner of social, political, and economic constructs.  Is the digital future a similar epoch?  Does it need new institutions?

Now that would be a domain shift to watch out for!