The Problem of Unpaid Work?

Just a thought …

Carl Benedikt Frey [Financial Times this morning] has me thinking.  How unwise of him!

An economy is the aggregate of all the work we want done.   It is not simply what we measure.  This is not revelatory, but it is becoming more important to think about the divergence between the measured and the total.

Why?

The steady rise of digital methods has allowed traditional industrial era activities to be deconstructed with the result that some things that were once included in the measured economy are now unmeasured.  Where once a customer was interacting across a very tangible boundary with a business provider in order to enter into a transaction, that same activity has now been unbundled and the customer has become a de facto and ad hoc employee because of the rise of “self-service”

Self-service is unpaid and thus unmeasured work.  This is well known.  The difference in our contemporary economy is that the portion of unpaid work is rising quickly as businesses offload activities onto their customers.   This is often presented as a “convenience”.  It is also a way to reduce costs because the work offloaded is no longer on the books of the business, but the transaction remains fully counted.  Productivity appears to rise as a consequence.

But its the economy any more efficient?

Only within the rules of the measured economy.   The amount of work needed to get the transaction complete remains largely the same.  The differences are who does the work, and is the work properly accounted for.  

As this process gathers momentum — and the rise of artificial intelligence is likely to intensify and accelerate it — we are all becoming a bifurcated participant in the aggregate of work.  We are both employed and on the books, and self-employed and off the books.  The balance is steadily tilting towards the latter.

As we slide ever more into a more service oriented economy this becomes of greater importance.  Service is, almost by definition, a lower productivity activity.  It has echoes of a one-on-one relationship between the provider and the consumer.  It is not a mass production and mass consumption activity the way large-scale manufacturing is.  So economies of scale are hard, if not impossible, to extract in a service activity.  

Adding machinery in a manufacturing process allows a worker to produce more.  Per-person or per-hour output rises.  The benefit of that extra output is contained within the boundaries of the business and thus measured as greater productivity.  The hope is that the result of that productivity gain shows up as higher wages for the newly empowered employee, lower prices for the consumer, and/or higher profit for the business.  In any case it shows up somewhere.

Automating a service is, as I mentioned above, different.  It introduces self-service and some of the work simply disappears from the measured economy.  It might till show up as profit for the business, or as lower prices for the consumer, and even as higher wages for employees, but the actual cost of the service is nowhere fully reflected.  Indeed, any potential lower price may well be offset by the additional work that the consumer takes on in order to fulfill the transaction.

Which, sort of, gets us back to transaction costs.

Digital methods, including AI, are lowering the marginal costs of executing a transaction ever closer to zero.  How far can this process go?  Deep into the firm itself?  How much of a firm’s activities can be thought of as a “service”?  Human Resources?  Marketing?  Accounting?  Executive management? Etc.  Are they possible to shift, digitally, into a self-service category?  And what about all those unpaid consumers criss-crossing the firm’s boundaries in order to self-serve?  They are replacing traditional work and thus traditional employees, with all the consequences such a process produces.  Where, exactly, is the firm’s boundary?

More to the point:  what is the marginal value of a near zero-cost activity?  The firm can, and will, offload it onto the consumer, but the activity retains value since the transaction cannot occur without it.  Now that it is beyond the boundary of the measured economy how do we associate value with it?  Is a value-less activity truly without value if a transaction cannot be completed without it?  It depends on what we measure.  

Is an economy really more productive when I pour my own coffee rather than when someone pours it for me?  Yes, if we measure output per employee.  Doubtful if we measure output per consumer.

What, then, would be the residual transaction cost justification for the existence of a service based firm?   The fixed costs of technology?  The intellectual property?  

And how do we measure an AI infused economy where an ever greater part of activity is self-provided by consumers?  How do we know productivity is actually rising when a large part of the necessary work involved in production is unmeasured?

Is it any wonder that capitalists are scrambling to move us into an AI economy?  They can extract profits unencumbered by employees — their workforce is one and the same as their customer base. 

Is this why modern distortions in the distribution of wealth are even greater than in past “gilded ages”?  And what does it mean for society at large if we are increasingly the unpaid employees of mega-sized technology firms that collect and concentrate the profits generated by all that unpaid work?

Is democracy tenable under such conditions?  Is that why so many of our technologically minded business leaders seem so disinclined to support democracy?  

How do we get that unpaid work back on the books?  Or, how do we get paid for what we contribute to the profits of those technology based businesses?

Just a few thoughts  …