Note: This is a guest post by Neil Kay. It is part of a chapter that he is writing for a book that I am editing with David Rooney and Greg Hearn called Handbook of the Knowledge Economy, volume 2. We’ll post Neil’s chapter as he writes it over the next few weeks. I’ll do the same with mine, which is seriously overdue too. – Tim

COMMENT FROM NEIL KAY
I puzzled some UQ MBAs in a guest seminar last year by telling them that I occasionally tried to become a better teacher by reading Economics textbooks upside down. While I would very much welcome comments on all the half dozen or so threads I plan to post for my article, I really would especially appreciate feedback on this thread as to whether it is on to something, or whether I should just keep taking the tablets.

-Neil

The complexity of economics and the Paradox of Mankiw

This is not a linear narrative and it is work in progress – I am going to jump to Thread 4 (or it might turn out to be Thread 5) in the story I told in my first post (“What does a knowledge economy actually look like?”). In that post I noted what I thought Mankiw lacked. In fact, having taught introductory first year economics using Mankiw’s book, I see it as being itself a superb example of how to succeed in a knowledge economy. For example, the standard monopoly problem in textbooks invokes all of the well-known AVC, AFC, AC, MC, MR and AR curves. Only certain curves have relationships between them or at least relationships that matter for pedagogic purposes (for example: AR and AC for profit; MC and MR for profit maximizing). The problem for a beginning student is that they have no mental map to tell them which pairs or combinations of curves have relationships that are significant in economic terms, and which do not (which is why many students often try valiantly in their exercises to get as many curves as possible intersecting at one point in the hope that at least some of the intersections reflect valid relationships).

If you want to experience an approximation to this tabula rasa effect yourself, turn an Economics textbook upside down and look at individual diagrams. Even experienced economists who do this tend to see a bewildering maze of lines, which is close to how a beginning student sees it.

One way to think of such curves is as members of a network in which each member may (or may not) have relationships (links) with other members of the network. For the beginning student lacking filters to sort out which of these are significant relationships, one measure of the complexity of the diagram (or network) is the total number of potential or possible relationships between curves in the diagram (or links between members of this network). For a monopoly diagram with all six curves (AVC, AFC, AC, MC, MR and AR) the total number of potential relationships or links1 between members of this network is 15 (or 5+4+3+2+1).

But if the monopoly diagram is split into two diagrams for the beginning student, say one with AVC, AFC, AC, MC and one with AC, MC, MR and AR, then the total number of potential relationships between curves (links between members of the network) in each diagram reduces to 6 (or 3+2+1).  That is a total of 12 potential relationships for the two 4-curve diagrams which in total is still less complex for the beginning student than the single 6-curve diagram. An experienced teacher knows that these two diagrams contain most of the relationships they want to talk about, which is why experienced teachers tend to use at least two diagrams here instead of one, at least to begin with.

I have a theory that if you used such a measure to assess the complexity level of first year Economics lecturers’ PowerPoints that they would inversely correlate with teaching evaluations – but that has not been subject to empirical testing.

Which is why Mankiw is a very good teacher and his textbook is a best seller. I do not know if Mankiw used a similar rule of thumb intuitively or explicitly, but his text is constructed as if he did. Most standard introductory texts are not shy about inflicting diagrams or figures with very high levels of network complexity on beginning students. But Mankiw has as few curves as possible in his figures, usually no more than 3 or 4. Indeed, in my edition there are only four figures with 5 curves and no figures with 6 curves or more. So here is another paradox; at the same time that Mankiw is doing the standard first year Economics textbook thing of describing “the economy” as one largely devoid of knowledge activities, his text may be regarded as indicative of best practice knowledge activity in a knowledge economy. The paradox of Mankiw is that in the economy he describes, his book would not exist.

Note: This is a guest post by Neil Kay.  It is the outline of a chapter that he is writing for a book that I am editing with David Rooney and Greg Hearn called Handbook of the Knowledge Economy, volume 2.  We’ll post Neil’s chapter as he writes it over the next few weeks.  I’ll do the same with mine, which is seriously overdue too. – Tim

When researchers write about the knowledge economy, they usually write about knowledge activities such as R&D, education, ICT etc, but what does – or what would – a knowledge economy look like? Describing a knowledge economy in terms of its parts is rather reductionist, rather like describing an elephant in terms of its trunk, tusks, legs etc.  So we shall try to see here what a knowledge economy would look like. First, back to basics – what is an economy?

‘There is no mystery about what an “economy” is … an economy is just a group of people interacting with one another as they go about their lives  (Mankiw, 1998, p.4)

That is not entirely satisfactory as a definition of an economy since it tells us what elements in the economy do (they interact) rather than what an economy is – Mankiw is describing process rather than content, “how” rather than “what”.  We can say that if it is about interactions of individuals it must be driven by the activities of these individuals. But what are these activities?  Mankiw answers the question later in the chapter (1998, p.21) where he introduces the standard circular flow model in which the activities are the production and consumption of goods and services.  Mankiw also gives the fictional example of Defoe’s Robinson Crusoe

stranded alone on a desert island as a basic example of an economy. Crusoe “catches his own fish, grows his own vegetables, and makes his own clothes. We can think of Crusoe’s activities – his production and consumption of fish, vegetables and clothing – as being as simple economy” (Mankiw, 1998, p. 519)

So if we can take “the economy” as being characterized by production and consumption activities, then it would seem logical to describe “the knowledge economy” as characterized by the production and consumption of knowledge activities. This in turn raises the question of what can be characterized as “knowledge activities”. It is perhaps easier to identify knowledge activities when we see them (e.g. lecturing, researching, design) than to give a simple definition that helps us distinguish knowledge activities from non-knowledge activities (and by implication, the knowledge economy from the non-knowledge economy). It is not a trivial problem; after all if we cannot say what a non-knowledge economy looks like, how can we expect to make progress with the concept of a knowledge economy?

If we are to make headway in terms of what characterizes knowledge activities, we have to establish what it is that is distinctive and characteristic about knowledge activities compared to other activities in the economic system.  In this respect, Kenneth Arrow’s paradox of the demand for information may be a good starting point:

“There is a fundamental paradox in the determination of demand for information; its value for the purchaser is not known until he has the information, but then he has in effect acquired it without cost.” (Arrow 1971, p.152).

I have been familiar with Arrow’s “paradox” since I was a student, but it is only now when I thought it might be useful in helping establish some definitive characteristics of “knowledge activity” that I noted a possible asymmetry here.  As Mankiw notes “the terms supply and demand refer to the behavior of people as they interact with one another in markets” (1998, p.62, italics in original)    If there is a fundamental paradox in the demand side of the market for information, would it not be reasonable to expect a corresponding paradox on the supply side?  I believe there is indeed a parallel paradox lurking on the other side of the Marshallian scissors, and that this is what it looks like;

“There is a fundamental paradox in the determination of supply of information; its cost to the supplier is not known until he has produced the information, but then he can produce it without additional cost.”

The “additional cost” aspect relates to the well known principle that information and knowledge may display public good characteristics in that the marginal cost of supplying it to an additional user may be zero, once it has been acquired.

Arrow outlined his demand-side paradox specifically in relation to inventive activity.  In such cases potential purchasers may not be able to place a value on individual R&D projects ex ante (Arrow’s paradox), while those undertaking the R&D face problems in specifying what the output will look like in advance, how much it will cost to obtain, or both (supply-side paradox).  If we paste these together, it suggests the defining characteristic of a knowledge activity is that it is not possible to specify in advance what its output will look like and/or what its value or cost will be.  The design of the Sydney Opera House displayed these characteristics to the full with both output (design) and cost calculations undergoing multiple and radical revisions in the course of the project.

If we now have a working definition of a knowledge activity, this may in turn help us establish the characteristics of a knowledge economy.  But crucially it should also help us establish what  not a knowledge economy  – for example, none of the economic activities described in Mankiw’s Robinson Crusoe economy such as catching fish, growing vegetables and making clothes are described and interpreted as knowledge activities.  It is true that Mankiw identifies technological knowledge as an integral part of economies including Crusoe’s, (1998, pp. 520, 523-24) but that is old or accumulated knowledge, it is not new knowledge acquired in the process of undertaking the activity in question. In standard textbook description of Crusoe-type economies such as Mankiw’s, Crusoe has all the knowledge required to function efficiently as both producer and consumer, there are no knowledge activities in which it is not possible to specify in advance what their output will look like and/or what their value or cost will be.  Mankiw’s Robinson Crusoe economy is not a knowledge economy.

But suppose instead we go back to original sources and observe what Defoe’s depiction of a Robinson Crusoe economy actually looked like.

Now this is where things get interesting and instead of pursuing this now, I could signpost some threads that could be pursued in the rest of the article –

Thread 1: In reality – which here means, rather paradoxically, in fiction – Defoe’s Robinson Crusoe could more properly be characterized as a knowledge economy in which Crusoe spent the bulk of his time engaged in various knowledge activities such as exploration, design, hunting, foraging, inventing, writing (his journal), signaling (for help) and teaching (Man Friday).

Thread 2:  Activities such as hunting and foraging in hunter-gatherer societies can be seen as knowledge activities, so were these early knowledge economies?

Thread 3: strip away knowledge activities from modern institutions and economies and you are not left with much. For example, Nike is entirely made up of knowledge activities such as design and marketing. So will be an interesting experiment to strip off layers of knowledge activities and see what we are left with.

Thread 4: knowledge activities can involve decision processes which take place over a prolonged period of time e.g. an R&D project, the design process for the Sydney Opera House. But in standard textbook (neoclassical) economics, a decision is not a process of generating knowledge but a choice at a point in time based on accumulated and sufficient knowledge.

Thread 5: so where have we finished up? One possible conclusion is that when we talk about “the economy” what can be described as “the knowledge economy” is a better (fuller, more accurate) representation of “the economy” than in the standard textbook models. If so the way forward is not to find new ways and tools to talk about “the knowledge economy”, the way forward is to replace inappropriate models of the economy in the economic textbooks with more accurate representations of the way “the economy” does actually work. And if that might seem an overly ambitious goal in the light of intellectual inertia, at least we should be aware of the limitations of the standard textbook tools in describing a type of economy which may not exist – and which may never have existed.

But I have not written the conclusion yet – writing this article is after all a knowledge activity in which the final output is uncertain

Arrow, K. (1971) Essays in the Theory of Risk Bearing, Chicago Ill., Markham

Defoe, D. (1995) Robinson Crusoe, London, Wordsworth classics.

Mankiw, N. G. (1998) Principles of Economics, Chicago, Dryden Press.