THE EVOLUTION OF DICK NELSON

remarks of

Sidney G. Winter

delivered at the Nelson Fest, New York, NY

October 13, 2000

It’s great to see you all here. I had barely walked in the door and saw all these wonderful and familiar faces, and I thought, this is just the sort of crowd that Dick Nelson can attract.

If I am going to discuss the Evolution of Dick Nelson, I suppose I should begin somewhere near the beginning. In fact, however, I know only a few things about Dick’s childhood and how his career might have been shaped by influences from that time. I do know that he lived in Washington as a child and became a loyal fan of the Washington Redskins, a fact that is still appreciated in our household.

There is one important thing that I did infer from my observation of him and from some comparison with my own introspection. Dick and I were both privileged to be members of the first generation of young fans of A.A Milne’s books about Pooh bear; Dick in fact was born only four years after the publication of Winnie- the- Pooh and two years after The House at Pooh Corner. (The fact that the Pooh books were of such comparatively recent origin did not actually come to my attention until I was middle-aged, and it came as quite a surprise – I had always had the impression that they were a few hundred years old, perhaps Elizabethan.) I am sure that the Pooh books were an important influence on his intellectual trajectory in ways both conscious and unconscious.

As evidence of the conscious type of influence, I recall a period when Dick frequently warned against the danger of falling into a Heffalump trap for Pooh-bears. This phrase provided a powerful metaphor for the dangers of muddled thinking and writing.

In case some of you are insufficiently familiar with the classics to appreciate just how powerful this metaphor is, let me sketch it for you. When a Pooh bear falls into a trap planned for Heffalumps, the result is not pretty. He generally winds up with a jar stuck on his head and then rushes around frantically bumping into things, and getting absolutely no place. Tellingly, Pooh bear was a party to the creation of the very Heffalump trap into which he later fell!

So please, all of you, try to avoid creating Heffalump traps as you craft those papers. And if you do dig the trap, at least take care not to fall into it yourself. If you do, Dick is sure to point it out.

Another thing I noticed is that somebody must have taught Dick the value of a dollar – and that must have been when the value of a dollar was eight or ten times what it is now. But here is the really important thing: Nobody ever taught Dick that Time is Money! And that omission, friends, has had wonderful consequences. It is God’s gift to us all.

Much of my observation of Dick’s evolution obviously derives from the four periods when I was privileged to have him as a colleague – two phases at RAND in the 60s, with a stint on the staff of the Kennedy CEA in between. Then there was the long period when we were both at Yale in the 70s and 80s, when we published papers together and finished our book, and also taught a couple of courses together.

In between RAND II and Yale, I was at Michigan while Dick was at Yale. We saw a lot of each other, thanks to the collaboration that was then building in momentum.

I think it was on one of my visits to New Haven, staying with the Nelsons in Hamden, that I first encountered Dick’s ritual forecasting of spring. My sense is that this ritual may have owed its origin to nostalgia for the warm sunshine of Santa Monica. Dick typically pronounced a week or two after the Groundhog, and he was a good deal more optimistic than the Groundhog on the average. Dick’s preferred indicator of spring was not his shadow, but the little snowdrop flowers. He would spot these peeking out from under a large snow bank and declare that spring was on its way. It was something of a feat of selective attention to spot the flowers against the snow bank, but Dick always proved right in the end.

Part of the evolution of Dick Nelson is obviously closely connected to the evolution of the Nelson-Winter book, An Evolutionary Theory of Economic Change, and it is obviously natural for me to tell the story from that perspective. Indeed, if we just take the chronological measure of that part we find that it occupies a full thirteen years – and that’s just counting the history that’s quite properly part of the undertaking that ultimately produced the book. There was about a decade of pre-history before that, as I shall now explain.

I “got onto” the evolutionary approach around the beginning of 1959. It was most directly a result of reading Armen Alchian’s article on “Uncertainty, Evolution and Economic Theory”. But I perceived that valuable stimulus with eyes that saw, not a general approach to the economics of the firm and industry, and not a quasi-methodological argument about profit maximization a la Milton Friedman, but the answer to an immediate and troubling problem: it seemed to suggested a promising theoretical perspective on the empirical question that was supposed to be guiding my dissertation research. That question was “What determines the level of R&D spending undertaken by a corporation?”

Mind you, I had started on the theoretical part of that work in the accepted way, with a profit maximization analysis of optimal spending on a single project. It soon became clear, however, that overall R&D intensity represented an organizational commitment that was not closely related to any aggregate of project level decision and hence was not intimately related to any cost-benefit calculations at all. The decision was more an expression of an attitude, often expressed quantitatively in terms of a desired R&D to Sales ratio.

The NSF surveys on corporate R&D spending had just begun; I had managed the considerable feat of becoming a sworn special agent of the Census so as to get access to the firm-level data. I was going to do multiple regressions explaining research-to-sales ratios. These, mind you, were the kinds and sizes of multiple regressions you could do on an electro-mechanical calculator, a Friden to be exact. I always liked the rhythm you often got when you did a calculation on one of those Fridens: whumpa-whumpa .. whump-whump whumpa .. whumpa-whumpa-whumpa-WHUMP!

The first point to note here is that I was, at that stage, very close to the intellectual turf that Dick was on at that time. He was studying the economics of technical change at RAND in the project that was led by Burt Klein. His much-cited paper on “The Simple Economics of Basic Scientific Research” was published in the JPE in June 1959. His valuable survey article, The Economics of Invention: A Survey of the Literature,” had appeared in April in the Journal of Business. (I think there may be a message in the fact that those two early papers were published in journals associated with the University of Chicago. In fact there was another one in the JPE, “Increased Rents from Increased Costs.” Got any more like that, Dick? Not that I remember.) The picture of Nelson at the time is that of a young economist who took justifiable pride in his command of the standard tools of the discipline and his ability to do creative research with those tools.

I joined up with RAND in September ’59, actually moving to Santa Monica in December. By that time, my research had moved in a theoretical direction, and more specifically in a methodological direction. I was trying to make sense of what Alchian and Friedman were saying about the relationship between profit maximization and economic “natural selection.”

Even before I got to Santa Monica, Dick had assumed the role of fan and sounding board for my theoretical work on economic natural selection. But my dissertation research pretty much went on the back burner when I went to RAND and then to the CEA, and the project, now entirely theoretical, wasn’t done until 1963.

It was not until our shared RAND phase II that things really started to come together. As I now see it, five discernible streams of influence converged at that time. These were

1) Dick’s interests in technological change and economic growth; 2) my interests in firm decision making and the related methodological issue about profit maximization; 3) the achievements of the Carnegie School, especially as represented by the Cyert and March book; 4) A renewed interest in Schumpeter, stimulated in part by our RAND colleague and mentor, Burt Klein; 5) exposure to the thinking of Michael Polanyi, especially as represented by his 1962 book Personal Knowledge.

This combination pointed us in particular to the question of what was going when behind the obvious fact that firms knew how to do things. At some point, Dick or I or both of us came up with a typical kind of scholarly conceit, which was the idea that if we really understood the theoretical problem of characterizing a single simple technology we would be a long way down the road toward understanding that problem in general. Whoever started it, we both got hooked on it. The simple technology that somehow became the focus for the discussion was baking a cake. I’m not sure why that was; I’m not sure either of us ever actually baked a cake – although I can do a tasty rice pudding. We may have been adopting the familiar economist’s device “Assume a can opener.” Only in this case it was, “Assume we are baking a cake.”

Anyway, this line of thinking led to what I will call the cake paradigm. Its first published version was in Nelson, Peck and Kalachek, Technology, Economic Growth and Public Policy, (Brookings 1967).

“Generally, a technique or technology is not describable by a unique routine; usually there are options in the program. These options permit some choice of inputs and input proportions (a recipe may work with either whole or powdered eggs) and some flexibility with respect to operations (the eggs may be added before or after the sugar). The operations may be performed in different ways; for example, different degrees of mechanization may be employed (the mix may be beaten with a spoon, a hand beater or an electric beater). Some variation in output specification may be possible (such as the shape of the cake or the kind of frosting). (etc.) pp. 99-100.

In 1968 I produced an informal RAND paper called “Toward a Neo-Schumpeterian Theory of the Firm.” It partook of the cake paradigm introduced by Nelson, Peck and Kalachek and by that time much discusssed with Dick. Importantly, it pointed out that Baked Alaska is not just ice cream and cake. From the ingredients point of view, it pretty much is, but from a process point of view, and hence from a knowledge point of view, it is not ice cream and cake. In a footnote, the paper briefly explored the problem of describing the technology of beating cake batter in a conventional activity analysis framework that recognized very fine distinctions in the time dimension as well as batter characteristics. The point was, it’s cumbersome to do that. However, my point right here is to remark on the persistent importance of the cake paradigm, so let me fast-forward to to close to the present.

From a 1998 working paper written with Bhaven Sampat, “Making Sense of Institutions as a Factor Affecting Economic Performance,” we have the following:

“Complex routines almost always can be analytically broken down into a collection of subroutines. Thus the routine for making a cake involves subroutines like pour, mix, and bake. These operations often will require particular inputs, like flour, sugar and a stove.” (p. 29).

So as you see, the cake paradigm has been something of a constant in Dick’s intellectual trajectory.

----------

When we both left RAND in the summer of 1968, there was a definite sense of a useful conversation interrupted. We resumed it in 1969, and after a period of wandering in the wilderness we hit upon the “book via articles” plan that we ultimately followed.

I obviously cannot tell that 13 year story in any detail. It seemed interminable at the time, and I do not want to reproduce that sensation here this evening. As we point out in the preface to the book, we did take a good deal of pleasure in a series of family celebrations, often on Cape Cod, of the “completion” of the book – to some stage or other. There were a lot of stages to celebrate.

Let me just offer a few selected observations on what it’s like to collaborate with Dick -- an experience that of course many of you have shared. It’s wonderful, but it’s not all a bed of roses. For one thing, Dick hyphenates “neoclassical”. I don’t know how many of those hyphens I had to remove.

Jonathan Cole pointed out in his remarks this evening that Dick encourages us to “engage complexity.” That’s quite true, but that encouragement often takes the form of a tendency to characterize a lot of things as “rich and variegated.” It’s a nice phrase, but a co-author does have to set some limits.

For the benefit of those of you who may be involved in collaborative authorship, I would like to mention one organizational principle that Dick and I devised and that proved quite helpful. It’s called the “no-reversion rule”. Suppose that you return our developing draft to me, after altering it because you found my expository prose unsatisfactory. Then, if I find your prose unsatisfactory, I’m not allowed to put my original language back! I have to try again. It’s a very helpful rule.

As this may suggest, Dick and I didn’t always find it easy to reach agreement on what we were saying and how we should say it. But we managed to see the whole thing through. And time and again, if enthusiasm started to falter and progress stalled, Dick provided the key energy to get things moving again.

------------

One of the ideas that we promote as evolutionary economists is the value of trying to identify the persistent traits of business firms, particularly the traits that underlie long records of success, or that particularly define the relationship of the firm to its niche in the economic system. This same sort of analysis can also help us to understand the key patterns that underlie a great scholarly career such as the one that we are celebrating this evening. Let me seize the opportunity afforded by my privileged place in the program and put in place some basic observations about Dick’s approach to the scholarly life. (Actually, speakers who occupied an even more privileged place have already mentioned some of them.) I am sure that others will want to reiterate, amplify and augment these points.

Three particular traits stand out for me. Although they are distinguishable, I think you will recognize that they are really all of a piece, or better, that they are just views produced by throwing different lights on the same singular phenomenon.

The first is a probing, playful mind, perhaps already illustrated by his enthusiasm for the cake paradigm. Dick Nelson with an idea is like a puppy with a favorite toy. He will sniff it and chew on it. He will pick it up in his teeth and give it a good shake to see what it is made off. He will back off a few paces and contemplate it – and then POUNCE; he’s on it again. If you try to hide it, he will sniff it out; try to toss it beyond reach and he will bring it back and look accusingly at you. In all this, he will be undeterred by the fact that the idea was sort of messy to being with and may be getting more so as a result of all the attention.

I have long thought that a certain kind of playfulness is indispensable when you are trying to discern the order in nature’s apparent chaos. If you take the whole thing too seriously, you quickly become discouraged by the difficulty of the task. For an example of this kind of playful spirit I give you Dick’s rendering of how knowledge guides search in R&D activity, from his 1984 paper in the QJE, “The Role of Knowledge in R&D Efficiency”:

“Stronger knowledge might lead to ability to focus the search more finely. Thus it might be learned that within the set of ‘blue’ techniques which was already known to be rich relative to the set of ‘yellow’ techniques, the blue striped subset is particularly rich. Alternatively, it might be discovered that, while the set of yellow techniques is a relatively poor one, the small subset of striped yellow techniques is unusually rich. R&D attention then might jump from the blue region to the striped subregion of the yellow region. Transistors are substituted for vacuum tubes, and subsequent R&D on amplifying devices focuses exclusively on semiconductors.” (p. 156 in Sources of Economic Growth.)

Of course, he probably could have chosen a more pretentious kind of abstraction than his colors and stripes, and conveyed those thoughts with set-theoretic notation and Venn diagrams and such. But he didn’t.

The second trait I want to remark upon is Dick’s seemingly inexhaustible fund of time and energy to engage other peoples’ work and comment seriously on it. How many acknowledgment footnotes are there that mention the helpful comments of Dick Nelson? They are like the stars in the sky. And you may know from experience that Dick’s comments are typically far more weighty than the average comment that warrants such acknowledgment.

The productivity of a scholarly community depends fundamentally on the adherence to “informal norms of reciprocity” in taking each other’s contributions seriously. The “informal” part in that sociological phrase means that nobody actually keeps the books on all these exchanges. If somebody did, we would be confronted with the fact that Dick is an enormous net creditor of the system, and our community would be convulsed with policy disputes about how to deal with the Nelson shortage. As it is, it usually doesn’t even seem like there is a Nelson shortage--which is quite amazing, and I don’t claim to understand it.

The third trait is Dick’s seemingly inexhaustible patience with the kind of meandering , frustrating discussion that typically takes place when people with quite different perspectives and enthusiasms are just learning to talk to each other. To an insensitive ear (of which I have at least one) it often sounds, frankly, like the conversation isn’t going any place. Yet Dick actually goes to a lot of trouble to convene such discussions. His willingness to do that and suffer the consequences has been a key element in his talent for producing distinguished multi-author volumes like Government and Technical Advance,

National Innovation Systems, and the recent one edited with Dave Mowery, Sources of Industrial Leadership. The sense of internal coherence one gets from those books, the underlying agreement among diverse authors on the questions that need to be addressed, is no accident. It is the fruit of a great deal of superficially fruitless discussion of the kind I referred to. As I said, Dick has endless patience for that sort of thing. With quiet optimism he keeps at the task of picking out the tiny flecks of gold in all that verbal mud and sand. Sometimes the thought crosses my mind that he actually enjoys the work, but that notion is so alien to me as to be quite discomfiting, and I have to put it aside.

There is one other Nelson trait, not so closely linked to the first triumvirate, that certainly must be mentioned. This is his moderation, his sophisticated sense of balance. On questions of public policy as on questions of economic doctrine, you will rarely find Dick uninformed about the arguments offered for positions contrary to his own. If he is uninformed, you will find him eager to learn. If the views of the opponents are worthy of respect, he will respect them. If the views are very dubious but the opponents themselves are worthy of respect, he will deal respectfully and gently with them. Sometimes, in the latter case, you get from Dick a kind of saintly sense of patience and resignation: “Forgive them, they know not what they do.” (Actually, in most cases what they do is called neoclassical economics, and they know that all too well.)

Happily, circumstances do arise occasionally when none of the above obtain. We value those occasions because they remind us that Dick’s moderation is not merely a matter of deeply ingrained routine; there is an element of deliberate choice involved. Dick does sometimes deliver very pointed remarks, comments that are delivered in the familiar modulated tones but translate very easily into something like “It is almost impossible to imagine an ignorance so profound as to yield up the sort of statement you just made.” We treasure these occasions for the rare jewels they are. Needless to say, it’s a good bet that the statement was even worse than Dick implied.

---------

Every now and then I make an effort to set down the central content of the evolutionary approach in a short list of propositions. Giovanni Dosi and I do a pretty good job of this in a paper to appear soon in the volume in honor of Richard Cyert. But every time I try this it comes out just a little different. In winding up my remarks this evening, I want to give you this evening’s version of the immutable principles of evolutionary economics. I start with the three methodological ones.

1. Realism! It may not be a necessity for good theory, but it is often a virtue at least at the prevailing margin. There is no need to take off one head and put on another one when you step reading the business page and start doing economics.

2. Dynamics first! To impose on dynamic theory the burden of supporting a pre-existing static equilibrium analysis, is essentially to put on blinders, making it inevitable that obviously significant issues will be overlooked.

3. No free calculation! It is an abiding scandal that the self-proclaimed science of scarcity routinely treats all forms of deliberation and information processing as free. This scandal reaches Monica-gate proportions in rational expectations and other sophisticated equilibrium concepts that implicitly endow each actor with the ability solve every actor’s problem many times over.

4. Firms are profit seeking! It is a true fact of nature that firms are typically profit seeking, but it is not a true fact of nature that they are typically profit maximizing. Profit maximization is a theorist’s crutch and ought to be abandoned when it is too stark to capture the reality of profit seeking or too cumbersome to permit analysis of any but the most extremely stylized models.

5. Innovation is always an option! One thing a profit-seeking firm can do rather than optimize over a given set of possibilities is to think of some new possibilities. Hence, every analysis of such optimizing behavior deserves an asterisk leading to a footnote that says: “Unless, of course, there’s a better idea.”

6. Firms are historical entities! They typically display pronounced inertial or quasi-genetic traits (e.g. scale/ routines) that are clearly persistent enough to shape their actions over interesting prediction periods. They ought to be represented that way in theory, positioned in model history the way real firms are positioned in real history.

7. Firms are repositories of productive knowledge! In most contemporary societies they are in fact the key repositories of technological and organizational knowledge and among the key agents of historical change. The storage and advance of knowledge, the maintenance and improvement of organizational capabilities, are complementary roles.

8. Progress is co-evolutionary! Technological and organizational innovation is generated by a variety of firm-level search processes. But firms do not search independently, they look to rivals, suppliers and customers for ideas, technologies and practices. And these firm and industry processes go forward in the context of a variety of public and private institutions and programs, which in turn are shaped by the firms. I could tell you that it’s really simpler than that, but That Would be Wrong.

9. Anything can happen for a while! As Schumpeter said, only “when things have had time to hammer logic into men” is it safe to assume that some level of rationality will characterize economic outcomes. Market discipline and economic natural selection constrain outcomes over time, but in the short run “anything can happen.”

Friends, there is a real program here. This is the way economics should go – at the very least, it is the way the economics of the firm, industry, technical change and economic growth should go. What’s more, it’s fun.

I have long recognized that these ideas, built as they are on many prior contributions, are a lot better than I am. They may even be a lot better than Dick. In fact, I actually think that they are better than all of us put together. If the ideas do not ultimately produce a major paradigm shift in economics, it will not be because the ideas are wrong or unfruitful. It will be because we have not pursued them with the energy and conviction they clearly deserve.

I do not intend a note of discord here. I want to thank so many of you who are present tonight for your contributions to making this program credible. I want to thank Dick particularly for his wonderful companionship on this great journey. But I do want to point out that there is a lot of work left to be done.

Do not be faint-hearted about this. Dick will help us do it.

Thank you.