secondary knowledge

/ / , , , , , , , , , /

More Complexity

Economists have been talking about complexity for a very long time.  This may surprise many of you given the state of mainstream economics, but it is true.  A good place to discover a preliminary history of complexity in economics would be the short volume edited by David Colander published in 2000.  The papers it contains were all presented at a History of Economics Society Conference in 1998.  Colander provides a good introduction and tries to put complexity into the context of economics since its inception.  There are times when, in my opinion, he errs a little too much in favor of the older methodologies as if he finds it difficult to shed old habits himself.  Nonetheless the papers in the volume cover a wide and interesting territory, including the similarities between biology and economics.

There is no point in presenting a summary of the various papers here, but perhaps it is worth mentioning that in Colander’s own contribution he evaluates some of the more well-known names in the history of economics and determines whether they would be more or less important were their standing simply built upon their understanding of the role of complexity.  The people coming under his scrutiny are a who’s-who of economic thought: Smith, Malthus, Ricardo, Mill, Marx, Walras, Marshall, Hayek, and Keynes all get a going over.  Happily for those us unconvinced by the dead-end that is called general equilibrium Walras gets a solid demotion, as does Ricardo.

In my last note I mentioned that economics has, traditionally, been heavily interested in what I called “a battle with scarcity”.  This seems to have confused people.  It is not exactly a novel idea on my part!  Robbins put it this way in 1932:

“The economist studies the disposal of scarce means. He is interested in the way different degrees of scarcity of different goods give rise to different ratios of valuation between them, and he is interested in the way in which changes in conditions of scarcity, whether coming from changes in ends or changes in means—from the demand side or the supply side—affect these ratios. Economics is a science which studies human behavior as a relationship between ends and scarce means which have alternative uses.”

This quote is familiar to you all I am sure.  So quite why my suggestion that economics has been interested in the battle with scarcity befuddles me.

Perhaps it is because we now live in an era of relative abundance and can spend more of our time concerning ourselves with the distribution of that abundance.  Perhaps, also, it is because we take for granted the enormous accumulation of practical knowledge over the past two centuries that have made this abundance possible.  There is a difference between thinking of the earth has being an abundant home for us all, and thinking about just how tractable that abundance is to us.  Obviously that part of the potential natural resources available in a practical sense is a function our ability to get at them.  This is why many of the narratives describing imminent depletion of resources have fallen foul of technological advance.  This is also why Malthus fell from grace despite having a very sophisticated view of the dynamism of an economy.  Indeed it is arguable that Malthus had a more sophisticated view of the complexity of reality than his famous rival, Ricardo.  What neither took into account was the consequence of constant learning and its impact on our ability to generate accessible wealth.  In other words our ability to overcome scarcity.

An added wrinkle, while I praise Malthus for his understanding of the dynamics of social change, is that his influence on Darwin remains controversial.  Economists took somewhat of a wrong turn when they put all their emphasis on competition as a driver of change, clearly reality includes cooperation as well, but they are not alone.  Altruism remains a tough subject for  evolutionary biologists to deal with — is it just a subtle form of selfishness?  In any case these words by Darwin could easily have been written by Malthus twenty years or so before him:

“all nature is at war, one organism with another, or with eternal nature.  Seeing the contented face of nature, this may at first be doubted; but reflection will inevitably prove it is true.”

Yes there are theoretical upper limits on how much we can take from the earth, but a more useful discussion is about our technological capability to approach that limit in anything remotely like a cost effective manner.  Theoretical and practical abundance are two separate things and the battle with scarcity is the battle to expand the space of the practical within the space of the theoretical.

This in no way diminishes concerns about ecological damage, it is simply a statement that economics has a historic concern with scarcity, which is an uncontroversial fact of history and not a shocking invention on my part.

Decision makers in an economy are constantly bumping into the reality of complexity, and since getting at a single “solution” or arriving at a single “equilibrium” is not reliably possible in a complex system they have a tendency to conflate it with a higher order of complicatedness.  I don’t blame them for this.  Indeed the essence of business strategy is to carve off a piece of reality and attempt to control it in order to make decision-making both possible and relevant.  So decision makers substitute a complicated system for a complex one and hope that the former is a good enough approximation to the latter such that their decisions yield the desired results.  Economists have gone down this path.  Rather than shift entirely into a complex mode of thought, they continue to model as if the economy is just extremely complicated.  Apparently Murray Gell-Mann’s saying that “the only model of a complex system is the system itself” has failed to penetrate far enough into economic thinking.

There are numerous features of a complex system that set it apart from a merely complicated one.  From my perspective the most important is that complicated systems rely on exogenous effects to induce change, whereas a complex system has within itself the ability to adapt and learn.  This is the single most appealing reason to think of an economy as a complex system.  Economies constantly learn.  Each iteration of an economy — each computation — allows it to build experience and allows it to include novelty in the next iteration.

This is why I split economic knowledge into two categories: primary and secondary.  Business activity centers on taking novelty and making it routine.  Routine knowledge, what I refer to as primary knowledge, is the grist of all business activity.  It is easily replicable, scalable, and controllable.  It is easily codified into business processes that can be managed towards a specified goal.  It is how we, to use Hidalgo’s term, embody knowledge into the products and services that form the core of economic transacting.  But primary knowledge has a great flaw: it depends on a static context.  By this I mean that the code assumes an unchanging environment in order to retain its efficacy.  There is no use in having the best buggy whip code when the future of transport is the internal combustion engine.  Secondary knowledge is that which allows us to adapt to such changes in context.  It is more heuristic and expensive to maintain.  It relies on experimentation and superior cognitive abilities.  It is the source of our ability to try new things in the face of the changed environment.

This simple division of knowledge is my adaptation of the work of people like the evolutionary psychologists Plotkin, Tooby, and Cosmides.

By the way I am not entirely sure that Gell-Mann meant that quote I gave above.  In his own description of algorithmic information content [AIC] he allows complex systems to be tractable to algorithms shorter than a description of the system itself.  He reserves the one-to-one description for chaotic systems.  In any case, the continuous rise of practical knowledge that underlies our rising prosperity can be looked at through the lens of AIC.  Early discoveries were in a space occupied by lower information content products and services.  They were easy to code, replicate, and produce.  Newer technologies gave us the ability to penetrate the space occupied by more complex systems with longer algorithms.  Each wave of discovery has allowed us to push further from the early simple systems.   But the ongoing goal of business is to find routines that make what appears to be complex into something more simple.  Inevitably that increases the interdependencies of the production process and moves us from a world of self-sufficiency into one of communal group dependency.

Which brings us back to Adam Smith and the division of labor, and the historic interest of economists in complexity.

Whatever that is.

 

 

/ / , , /

Speculation

Robert Locke’s excellent discussion of the different perspective presented by a tacit-knowledge rather than explicit-driven driven enquiry into economic matters prompts me to reprise my understanding of the purpose of a business firm.

Put briefly: a business firm exists to translate what is often called tacit knowledge in to what is called explicit knowledge. An alternative wording would be that firms take the ad hoc and various and make them into the codified and uniform. In my own thought I use the words “secondary knowledge” to refer to the kind of knowledge that tackles ad hoc or unexpected circumstances. I use the words “primary knowledge” to refer to the codified knowledge that allows us to tackle the usual and/or repetitive circumstances. It is phraseology I borrowed from evolutionary psychology.

Why do firms do this?

Because that’s how they extract value.

Firms succeed when they control a process in order to produce something. Such processes are constructed from a variety of roles and routines whereby knowledge is combined with energy and physical resources, with the outcome being something for sale. Since establishing a process is a risky proposition — it takes up both space and time and so is susceptible to uncertainty on two fronts — control becomes the key component. And, in order to make the process most profitable, firms want to establish maximum control over the outcome: products must be both qualitatively and quantitatively predictable. How best to do that? By eliminating the ad hoc and various and limiting production to the codified and uniform.

However, as all students of evolutionary theory will recognize, being too specific, too codified, and too uniform renders something highly risky in the face of uncertainty. An unpredicted event may make the output of even the most efficient process useless or obsolete. In biology such uncertainty might arrive in the form of environmental turmoil. In business it might arrive in the form of a new technology. Either way excessive specialization and excessive efficiency makes a process vulnerable to change.

So firms need to absorb novelty. They need to be able to adapt. This implies that there is survival value in the ad hoc and the various, since they are the source of the novelty and adaptation. But to wring out the efficiencies upon which profitability best thrives the new ideas and discoveries within the ad hoc and various must be reduced to code: reliable replication is the key to effective production.

The profit cycle then begins again: knowledge gleaned from confrontation with uncertainty is carefully reduced to code. The specific discovery is made general production. And along the way value is added because quality controlled production becomes possible in quantity.

Being able to deploy both primary and secondary knowledge is the only long term source of survival. In my view the goal of a firm is to survive, it is not to maximize profit since uncertainty rules out the ability to maximize. But to thrive a firm needs to generate sufficient profit to locate and benefit from novelty so it can then reduce that novelty to codified and thus controlled production.

And so on …

The advantages of primary knowledge are that it allows consistent and reliable replication; it is objective inasmuch as it exists outside of any one person [it is coded]; and it is cheap to maintain [it has low/no error rate]. It is thus the most efficient basis for process construction. As noted: these advantages come at the cost of potential failure in the face of change.

In contrast, secondary knowledge is highly variable; subjective inasmuch as it is unique to an individual; and it is expensive because it is both relatively scarce and inscrutable. It is also highly likely to involve redundancy since its components might have no application to any particular problem, but are, rather, available to be called upon to deal with change.

So the getting balance between highly efficient but vulnerable to change, and less efficient yet adaptable, processes is the goal of good management. Survival depends upon the latter. The former is the source of profit. Translation from one to the other is the purpose of the firm.

 

/ / , , , /

Notes On The Firm

When, for reasons too byzantine to recount here, I decided to re-engage with economics a couple of decades ago I did so through the prism of business. After all I had just finished a stint in banking and was thinking about the way in which the rise of digital technology would change the way in which business is conducted.

The reason for this point of re-entry was that I believed then, and still do, that economies are movements of information as much as, if not more than, anything else. So if digital technologies make information more tractable, abundant, and accessible then they must affect business. My earliest basic assumption was that a business firm is simply a system of thought whose objective is to protect and exploit its information advantage. This brings it into conflict with both its suppliers and its customers who are seeking to overcome the asymmetries of information that occur “naturally” in the landscape. I described the space that firms occupy as “operating space” that is a wedge preventing direct contact between customers [end users] and suppliers [or resources].

That was back in the mid 1990’s and, by and large, since then the disintermediation of operating space that we observe as a consequence of the “digitalization” of the economy has gathered pace.

The most obvious manifestation of this is the steady redesign of business itself. More and more business firms are reducing their operating space by jettisoning  what they consider to be “non-core” activities. The idea being to reduce their logical space only to that in which their information advantage resides, and to allow those aspects of their older form that represent no advantage to go elsewhere to be repurchased at lower cost if, and when, needed.

This, fundamentally, is what outsourcing is about. It is a recognition that the information advantage within certain activities has been eliminated by digital technology. Those activities are now better accumulated and managed outside of the firm, perhaps by other specialty firms, or even by individual workers who can act as contractors with many customer firms rather than being an employee captive within a single firm.

The greater availability of information, and its greater ease of communication is, in my opinion, the key to understanding most trends in business. It certainly underlies the radical reconstruction of our retail industry where the older model incorporating a physical presence in stores scattered around the country is being made obsolete by a more modern and more efficient digital marketing method coupled with a digitally enabled logistical infrastructure.

Key to understanding all this is having an idea of what the objective a business firm is.

In my mind a firm exists to exploit its information advantage. It must, therefore, pay particular attention to its boundaries, not in a physical sense, but in a logical and information sense.

To extract value from its information advantage a firm has to create and manage processes that produce whatever it is that it sells into the outside marketplace. And to secure the most [not the maximum since that is an impossibility] it seeks to be as efficient as it can be. Efficiency in a business sense then becomes a consequence of creating product knowledge, and making its production processes reliable, replicable, and predictable. This is why firms prefer to deploy what I call “primary knowledge” rather than “secondary knowledge”.

Primary knowledge is that which can be encoded in some form of instruction to create routines to produce a product or artifact. That artifact doesn’t have to be a physical thing, it could be something intangible, but the goal is that production is consistent and predictable. Each car, for instance, coming off a production line must conform to a plan. The plans embody the primary knowledge required to produce a car. The more rule-based the plan the more likely it is that production can be scaled to higher volumes; the more likely it is that each car will be standard; the more likely its is that errors are eliminated; and so on.

The problem with primary knowledge is that its efficacy relies on a stable environment. It is, after all, the basis of routine. Wherever a task can be made routine primary knowledge, that is codification in some form, will create the greatest value. This is why robots are superior to humans on production lines: they are less susceptible to error since their underlying code is more predictably followed. Humans have a greater tendency to err. Primary knowledge, because it is encoded, sits at the heart of the transition from the industrialization to digitization.

Let me repeat: business firms exist to exploit information. They do this by encoding what it takes to produce something. They translate knowledge into code. The more of their activities they can encode the more value they can squeeze from their information advantage. So businesses are “in the business” of creating routines. The more routine the better. Business creativity is thus almost a contradiction: it is in the discovery and implementation of routine. More particularly it in the discovery and implementation of  routines that either supplant older routines by virtue of greater efficiency, or which are novel and thus allow the production of new products.

With business firms recognized as code-making and thus efficiency-seeking devices the issues we all have with economics comes into sharper relief: economic theory is pointless in the face of the real world of information processing.

It’s the environment. Real world environments are riddled with uncertainty, emergence [you cannot predict the whole from the sum of its parts], computational intractability, and the simple fact that history is no certain guide to the future [the world is non-ergodic]. In other words, there is an ever changing context that cannot be predicted, calculated, or even easily envisioned.

Business deals with this by creating “alternative futures” called plans. These plans are simply efforts to project a modest amount of certainty into the future, knowing that such a future is unlikely to transpire exactly, but that the match between the prediction and that which transpires is sufficiently close for current decisions to be made.

Businesses need stable futures so that their codes remain relevant. The moment a business context changes the code risks becoming irrelevant. It is a question of adaptation. Even the best code can become useless if the instructions it embodies are no longer efficacious because the environment into which the product emerges is too different.

Which is why secondary knowledge exists. Secondary knowledge is best thought of as the ability to learn or adapt. It is secondary because it is vastly less efficient in stable environments. It is less efficient because it involves the maintenance of knowledge that might not be currently useful but which may be useful in some unknown future. Such knowledge is embodied not so much in code as in the persistence of sets of rules to apply to solve new problems. These rules may, of course be encoded, but their inefficiency resides in their current applicability: there is a cost in carrying around code that has no immediate application. So the ability to learn is an expensive and often unnecessary quality. It becomes vital in periods of rapid change, but is a luxury in times of relative stasis.

Thus firms prefer primary knowledge: it is less costly to maintain, more predictable, replicable, and reliable.

This leads us to another insight: business firms are devices for the translation of secondary knowledge into primary knowledge. They are not designed to, nor do they necessarily want to, learn. Learning is expensive.

This is why it is devilishly difficult for existing firms to innovate well. Innovation is a form of learning. It is an activity that sits uneasily within the codifying apparatus of a well-established business. Naturally, all businesses retain a capacity to learn by virtue of their human basis, but as logical entities they resist too much learning. Their entire reason for being is to seek efficiency through codification. Learning disrupts that efficiency seeking goal.

But, enough, these are notes to myself after all.

Summary: only in the context of an economy as the movement of/exploitation of information can we build a good theory of the firm.

Mainstream economics cannot do this because it cannot deal with radical uncertainty, non-ergodicity, emergence, or computational irreducibility. Indeed it refuses to acknowledge these facets of reality. These facets all have information in various forms as a common theme. Business, in contrast, must deal with the asymmetry of information. Indeed such asymmetry is why firms exist. So by looking at the way in which real firms behave and deal with information we can begin to understand how economies as a whole work.

Without asymmetries of information business as we know it becomes impossible to sustain. The distinction between production and consumption would blur, and the institutions of our modern economy would need radical replacement.

But that’s for another day.

 

/ / , , , , , , , , , / 0 comments

Human Capital – The Knowledge Dimension

The very words have a bureaucratic ring to them. At least to me. Nancy Folbre has written an interesting article about the ending of the Golden Age of Human Capital, her main point being that there has been a major change in the landscape for higher education. The value of a college degree is not what it used to be, and the cost of getting one is becoming prohibitive.

The second part of that statement – about the cost of education – is testimony to the shockingly irresponsible shambles that academia has become. The cost of providing an education is way out of line, affordability is declining rapidly, student debt shackles the young and reduces their ability to attain the creature comforts of the old middle class, and the future value of college is increasingly murky. Academics decry the cost of getting an education, but offer little by way of a remedy. The productivity within academia is a challenge to say the least, and odd ball solutions like opening up campuses of our universities in China solve nothing. Indeed they make matters worse by building the supply of graduates and thus eroding the return available for all graduates.

Folbre hints that there may have been a shift in the demand for educated workers, at least here in the west, with employers now focusing on more specific skills rather than the old fashioned generalists pumped out en masse by liberal arts colleges.

Let me add my voice to this discussion.

Corporate America is awash with highly educated bureaucrats. The demand for such folk rose strongly during the 1980’s as big businesses sought to upgrade the talent of their management teams. Prior to then it was not unusual to find a large corporation being run by a raft of ‘corporation men’ – invariably men – who had risen through the ranks and benefitted from employer specific training. Sometime around the early 1980’s in-house training was identified as being too expensive. Those programs thus became an early example of outsourcing, with the cost of management education being pushed onto the employee and off the corporate income statement. This was acceptable to employees as long as the returns to the cost they were now covering was significant enough, which, as we all know, they were. By the end of the 1990’s most employer specific management training programs had disappeared and had been replaced by minor technical training, and, sometimes, the subsidy of employees who sought their own education whilst at work.

This is, of course a generalization, but notice two convergent trends within this history.

First, the early 1980’s was also the onset of the single minded pursuit of shareholder value. Experience and education became costs to be cut or outsourced like any other cost. The drive for efficiency – to pump up returns on equity – drove corporate America to abandon many older methods and ways of doing business. Among those jettisoned notions was lifetime employment and the ‘corporate man’. Instead of providing training big businesses turned to the marketplace to hire already educated employees. This created a boom in demand for anyone with a higher education. Thus the emergence of the Golden Era for Human Capital that Fobre regrets the passing of.

Second, this surge in employment for educated people built the supply of able bureaucrats capable of seeking the very efficiencies that had caused the outsourcing of business education. The ranks of business swelled with accountants, lawyers, consultants, and other highly educated workers whose pay seemed immune to the stagnation witnessed by less educated workers as efficiency became ever more important as a source of profit. These ‘knowledge workers’, as they began to be called, were the ones leading the efficiency charge. They were the ones steeped in the latest management and economic doctrines, and they were the ones reaping the rewards from the slimming down and streamlining they led with such gusto.

It wasn’t the pursuit of innovation that drove the demand for knowledge workers, it was the drive for efficiency. Creativity was valued more for incremental tinkering with corporate costs, not for breakout product or process changes. Despite all the hype from management gurus, innovation was the lesser source of short term profit, by a long way, than was the search for efficiency. As long as costs could be cut the demand for bureaucrats willing and able to cut them was near insatiable.

Sooner or later, though, the fat had been trimmed and the demand for such bureaucrats tapered off. The newly lean corporations could manage themselves with fewer managers. The relentless pursuit of shareholder value brought the high cost of those bureaucrats into focus. Business decided it needs fewer. It began to outsource the activities they occupy. The drive for efficiency entered a new phase.

Which is where I think we are.

The value of those higher educations has diminished for many, leaving ever larger spoils for those who manage to remain. The top 1% of wage earners began to separate from the top 5%. The skilled bureaucrats began to feel the same heat that they applied to lesser educated workers.

Let me step back and add a wrinkle to the story.

In an economy there are two kinds of knowledge. What I call ‘primary’ and ‘secondary’. They are distinguished by the role they play rather than the content. This is different from the tacit/explicit dichotomy of Michael Polanyi in his 1966 book “The Tacit Dimension”, and which has become so popular amongst management and consulting theorists.

The two kinds of knowledge relate to the two impulses upon which an economy is based. Innovation creates the long haul rise in wealth that we all benefit from – and which the McCloskey book I commented on last week brings into sharp focus. Efficiency re-arranges the existing resources to reap greater rewards from them.

In my terms: primary knowledge is the source of efficiency; and secondary knowledge is the source of innovation.

Primary knowledge is embodied in the skills of workers doing repetitive tasks, in the capital machinery and technology that clutter the business world, and in the management methods that dictate the structure of production processes. Primary knowledge allows the pursuit of efficiency because it can be economized upon. It can be replicated exactly. It can be transferred. It can be coded. It can be compressed. It is the stuff of corporate process. These qualities allow for the reliable reproduction of product. Because they are so predictable and replicable, they allow for the minimization of cost and the maximization of profit – don’t forget that the pursuit of a maximum requires a stable and well known set of inputs for the computation. Primary knowledge needs known problems with known solutions. It assumes a future that mirrors the past so that its efficacy can endure.

Secondary knowledge is very different. It is embodied in the problem solving ability of workers and machines capable of ‘thinking’. It is how we deal with new problems and issues. It does not assume that the future is a mirror of the past, but that there exists a capability to innovate when presented with a need or an opportunity. It creates new solutions to new problems. But there is a downside: it cannot be economized upon since it exists as a vague potential up until the moment of problem solving. This, by the way, is why McCloskey can subtitle her book: ‘why economics can’t explain the modern world’ – standard economics deals only in a world of primary knowledge and maximization and has nothing to say about ad hoc problem solving. Secondary knowledge is also expensive since it requires a complex of available routines, thought processes, and associated information, much of which may never be called into action. It is idiosyncratic and personal, making it intractable to both to business, and to standard economic theory. Yet it is the source of the breakthroughs that constitute innovation.

This dichotomy adds, I think, a dimension to Folbre’s analysis. The surge in interest in human capital coincided not with a heightened interest in creativity or innovation – despite the claims at the time – but with the need expressed by corporations to translate their existing secondary into more efficient primary knowledge. In other words the drive to embody existing knowledge and ring the last ounce of profit from it produced a demand for bureaucrats to re-invent older products and processes rather than to invent new ones.

Once that drive had run its course, and as our economy had been compressed to produce the most profit, the demand for ever greater bureaucratic armies ebbed.

Interestingly, this narrative explains some of the macro trends that others attribute to structural problems in the economy. People like Tyler Cowan, for example, predict a diminished trajectory for GDP growth because our economy is not innovating as well as it once did. According to Cowan, our latest inventions have a far smaller impact on future wealth creation than did those of a century ago. This conclusion feeds into his standard right wing cry for freeing up enterprise and the reduction of social entitlement programs. He says we cannot afford those programs because of the diminished future, and if we want to move the growth curve back upwards we need to reduce government controls.

But my narrative produces a different interpretation: the cause of our diminished future resides in the private sector’s single minded pursuit of profit being extracted from efficiency rather than from innovation. We have succeeded mightily in squeezing profits from our current set of  ideas. But at the cost of thinking about and finding the innovations that build the future.

We have become overly bureaucratic, technocratic, and reliant on primary knowledge.

I think the Golden Age of Human Capital is yet to come. The age of problem solving, that is. Not the age of rote learning.