Created 10/7/1998
Go to Brad DeLong's
Home Page

Teaching | Writing | Career | Politics | Book Reviews | Information Economy | Economists | Multimedia | Students | Fine Print | Other | My Jobs

is always very welcome...

Rules, New and Old, for Tomorrow's Economy

Carl Shapiro and Hal Varian (1998), Information Rules: A Strategic Guide to the Network Economy (Cambridge: Harvard Business School Press: 087584863X).

Kevin Kelly (1998), New Rules for the New Economy: Ten Ways the Network Economy Is Changing Everything (London: Fourth Estate: 1857028716).


J. Bradford DeLong

Professor of Economics, University of California at Berkeley, and Research Associate, National Bureau of Economic Research

forthcoming in WorldLink: The Magazine of the World Economic Forum

Draft 2.1; October 7, 1998

I. Introduction

The past four centuries have seen one economic revolution after another: the commercial revolution, the industrial revolution--the first application of steam and mechanism--the so-called second industrial revolution of steel and chemicals, the third industrial revolution of electric motors and internal combustion engines.

Now we find ourselves in the middle of another economic revolution, this one centered some forty miles south of San Francisco, perhaps the fifth such economic revolution since 1600.) But what do we call it? This industrial revolution has too many names: is it fundamentally about silicon, or microprocessors, or computers, or telecommunications, or--to become abstract--"information," or (to become dizzyingly abstract) "the network"?

And even after we have decided what to call it, there remains the question of what to think about it. How much credence are we to give to claims that this particular economic revolution is creating a genuinely "new economy"?

We have had so many economic revolutions in the past several centuries. So an inevitable question to ask is: Is this "revolution" a genuinely "revolutionary" revolution--giving rise to a genuinely new situation, in which things will be fundamentally different than anything in the past? Or is this particular "revolution" just another more-or-less standard revolution--one that will fundamentally transform some industries, make some very large fortunes, enlarge human capabilities, enhance our material wealth, but at the end of the day leave us with an economy that still works in the same way? In their days Detroit as the home of mass production and Manchester as the home of cotton spinning were as "revolutionary" as Silicon Valley is today. Yet after the initial industrial and the mass-production "revolutions" were over, the economy looked much the same--albeit with much, much cheaper cloth in the first case and much, much cheaper vehicles in the second.

These two books resolve the first problem--that the current economic revolution has too many names--in the same way. The authors of both books have chosen subtitles that call what is coming into being the "network economy." This is a very important area of agreement. It means that both sets of authors have chosen to focus not on technology, but on social and economic organization. Thus these books have very little on C++ or Java as programming languages, on how the promise of quantum computing can extend the reach of Gordon Moore's law that silicon circuits (and costs) halve in size every eighteen months, on the interface of fiber optics to the copper last mile of telecommunication service to the home, or on the increasing scale of required investment in chip fabrication. Instead both books have a lot to say on how markets and firms will shift over the next generation or so, on what kinds of pricing strategies successful firms will pursue, and on how the changing technological underpinnings will change what kinds of firms will survive and flourish. One book offers "rules for the new economy"--rules for how firms should behave, that is. The second sells itself as a "strategic guide"--again, strategies for firms. These two books thus take what is in some sense the same cut, a social and economic organization cut, at the whole complex of issues surrounding Silicon Valley.

But there the similarity between these two books (both of them very good books) stops. For they give fundamentally different answers to the question of how "novel" is the network economy.

For Carl Shapiro and Hal Varian's Information Rules: A Strategic Guide to the Network Economy, our current economic revolution is a non-revolutionary revolution. They see much in common between the telecommunications and other industries reacting to technological change today and the telecommunications and other industries reacting to technological change a century ago. Thus to look forward Shapiro and Varian are very likely to look back. And they claim that you don't need a "new economics" to understand tomorrow's "network economy." After all, "[t]echnology changes. Economic laws do not. If you are struggling to comprehend what the Internet means... you can learn a great deal from the advent of the telephone system a hundred years ago" (Shapiro and Varian, pp. 1-2). All you need is "to see the really cool stuff" that was already in the old economics, but that was just not taught in undergraduate surveys (Shapiro and Varian, p. viii).

For Kevin Kelly's New Rules for the New Economy: Ten Ways the Network Economy Is Changing Everything, by contrast, this current economic revolution is creating a genuinely new economy--a "new economic order" rooted in the "distinctive economic logic of networks." As a result the "[e]conomists are baffled."

Now these differences may be as much differences of style and tone as substance and analysis. Kevin Kelly would be the first to argue that there are powerful and fruitful parallels between the economic revolution of today and past episodes of rapid innovation in narrow sectors. Kevin Kelly is a guy who sent a very good science-fiction writer (Neal Stephenson) around the world to chronicle the long-run impact of the "wizard-hacker-lords" of the second half of the nineteenth century who pioneered the use of electricity for long distance communication: that is taking Shapiro and Varian's advice to look back in order to look forward with a vengeance. Out of the other authorial team, Hal Varian was lured away from a comfortable life in the University of Michigan's Department of Economics to become the Dean of Berkeley's Library Science School--or rather, it was the Library Science School until Varian got control of it: now it is the School of Information Management and Systems in which you are more likely to find Macarthur Prize Fellows thinking about the crossroads where law meets technology than experts in the Dewey Decimal System.

Kevin Kelly is a professional visionary--yet at least sometimes he believes that circumstances change but laws do not, and that close study of similar episodes in the past will help illuminate the future. Hal Varian is a sober and professional economics professor--yet is visionary enough to to become the first Dean of the first educational institution of its kind.

Yet the two authorial teams would, I think, admit to sharing little because differences of style and tone do reflect and become differences in substance and analysis. If the two sets were ever on the same podium at the same time, sparks would probably fly. Kevin Kelly would begin by quoting Walter Wriston on the irrelevance of economics and economists. He would denounce traditional economists as focused on equilibrium and order when the nature of the network economy is "disequilibrium, fragmentation, uncertainty, churn" (Kelly, p. 151). He might conclude that: "[t]he dials on our economic dashboard have started spinning wildly, blinking and twittering as we head into new territory. It's possible the gauges are all broken, but it's much more likely that the world is turning upside down" (Kelly, p. 3).

Carl Shapiro might reply by denouncing those who merely extrapolate trends--"more decentralized, more organic, and more flexible... flat organizations and unlimited bandwidth"--rather than build models and so have no idea whether the trends they spot will in fact continue. Hal Varian would then make fun of "vocabulary for the sake of vocabulary," and conclude that what readers need are "... models, not trends; concepts, not vocabulary; and analysis, not analogies" (p. 18).

So let me note and emphasize the differences between the books--because noting and emphasizing their differences will help me make my argument more conveniently. For I wish to offer a qualified allegiance to Shapiro and Varian. The network economy that is being born in Silicon Valley does have much more that is old about it than new. This economic revolution is thus--mostly--a non-revolutionary revolution, that should leave the fundamental structure of the economy with little change.

But my allegiance to Shapiro and Varian cannot be complete and absolute, but must be qualified. Where the two books agree, I find Shapiro and Varian's discussions to be much more comprehensible and correct than Kelly's. (It would, after all, be a great surprise if I did not: I am like Shapiro and Varian a professional economics professor; I am not like Kelly who is a professional visionary.) But Kelly's field of vision seems to me to be somewhat broader: he sees interesting things that escape notice in Shapiro and Varian's strategic guide to the network economy, things that may in the end wind up making our economy truly new.

II. Business Strategy in the Network Economy

The very first point that Shapiro and Varian's Information Rules makes on the very first page of its very first chapter is that we have a long history of rapid technological advance. Thus we have a lot of cases to study when we try to think about how rapid technological progress affects the economy. And even though technologies change, Shapiro and Varian claim that the underlying economic principles do not--and it is the underlying economic principles that determine success and failure.

In their view, you only have to grasp three sets of economic principles in order to grasp the key information rules. The first is that information goods are "experience goods" in which consumer value varies extremely widely. The second is that information goods have a "rather unusual" structure of costs. And the third is that the network itself is a source of powerful externalities on the demand side.

Unpacking these three sets of economic principles--first explaining what they mean, and then applying them--is the heart of the book.


A. Demand

How much should I be willing to pay for the privilege of reading the _Wall Street Journal_? Some people find it very valuable. Others find it a waste of time--tendentious and short-sighted editorials, and news stories that are either irrelevant or that tell them things they learned through the gossip vine a week ago. There's no way to really tell which is the case without reading the Wall Street Journal for an extended period of time. And things change: when I lived in Washington, D.C. I found the Journal much less valuable than I do now that I live in California.

Shapiro and Varian classify those who sell the Wall Street Journal--classify practically everyone in the business of purveying information--as having the difficult task of trying to sell something that is an "experience good." You can't show people how valuable your product will be to them unless you let them experience it. But once they have experienced it--heard what you have to say, learned the piece of information you have to sell--they then have no reason to buy it. Why should they pay for something that they already know? But if a seller doesn't let people know what he or she has to sell, how will anyone ever learn that they would benefit from buying it? In Shapiro and Varian's eyes, this is perhaps *the* fundamental problem facing businesses in the network economy: "[t]he tension between giving away your information--to let people know what you have to offer--and charging them for it to recover your costs" (Shapiro and Varian, p. 6). Either way you capture only a very small part of your potential market as paying customers.

Rescue comes from another key fact about demand for information goods: the value people place upon information--how much they are willing to pay--varies especially widely across individuals and also across time. A very large chunk of the aggregate willingness-to-pay of a firm's potential market is contained in a relatively small chunk of its potential audience. Thus Shapiro and Varian arrive at the key strategic imperative in managing demand: segment your market. Sell at a premium price to those customers who are willing to pay a lot, those who value your information good highly. And sell at a steep discount price to those casual customers with a low willingness to pay--or perhaps give your product away free to the masses in the belief that wide distribution is the best way to pull more premium-price customers into your market.

There are two ways to sell to high willingness-to-pay customers at a premium price. The first is to sell them a premium product: Reuters succeeds at charging a premium price for the news flow by doing a better job at sorting the news its clients are interested in from the news its clients aren't interested in (Shapiro and Varian, p. 29). Book publishers have long used the difference between cloth and paper covers as a way of separating out those with a high willingness to pay (who buy the hardback because they want the book *now*) from the rest (who wait for the paperback). Software publishers release free and premium versions of their products, thinking that it is worth it to distribute the free version of the file-decompression program to twenty casual users if by so doing they can make one serious high-margin user aware of the extra features of their premium product.

The second way is to try to not offer your premium customers the opportunity to buy at the lower price at all. Right now the grocery store where I shop most often--Safeway--is trying to train me to always hand the cashier my Safeway Club Card so that Safeway can get and keep a complete record of my purchases. I expect that five years down the road I will be offered individualized me-only coupons for Safeway products for which their model believes that a small price reduction will induce more purchases, and no coupons for products which Safeway thinks I will buy anyway.

Shapiro and Varian see the coming network economy as both creating the pressure for firms to segment their markets in every way possible--for no firm can afford to do a worse job than its competitors at charging premium customers a premium price--while also bringing with it the information-processing power to do a professional job at market segmentation. For if a firm doesn't find a way to charge premium customers a premium price, it will have nothing to sell but generic information to low willingness-to-pay customers, and on the net "the generic information...--information commodities such as phone numbers, news stories, stock prices, maps, and directories--[is] simply selling at marginal cost: zero" (p. 24).


B. Supply

Businesses that do not figure out how to segment their markets will not survive because the cost structure of the network economy makes it impossible for market structure to even approximate the many-producer, perfect-competition benchmark standard in economics textbooks. In most industries minimum efficient scale is not that large compared to the size of the market. There are no extraordinary efficiencies from large size that create low cost, and larger entities are more likely to find their costs running high as they trip over their own bureaucratic feet. But information goods are different. There are very high--and sunk--set-up costs: it costs a lot to make the first CD-ROM disk for a product, or to write the program. Moreover, once you have sunk the initial cost it cannot be recovered: it is not like a building that can be refitted and rented out for another purpose. There are very, very low marginal costs to making another copy of the CD or another copy of the program--if you are clever enough to figure out how to sell your commodity over the internet, your buyers will even pay the reproduction cost for you. And there is no capacity constraint: no reason why a firm cannot expand production to meet demand within the course of a day.

In some other industries there are no large economies of scale: so perfect competition in the economist's sense is achievable.

In still other industries fixed costs are not sunk. Hence producers that are losing money but still covering variable costs can exit the industry and turn their fixed assets to another use. This process of exit helps keep prices above their marginal cost floor even in industries with large economies of scale: even though economies of scale are inconsistent with textbook perfect competition, it still works almost like a competitive market because it is contestable by new entrants who can test the waters for a while.

In still other industries economies of scale can be high, and fixed costs can be sunk, but some form of oligarchic competition can still be maintained. Boeing and Airbus Industrie have high fixed costs and enormous economies of scale, but neither could ramp up production to take over the other's market share in any period much short of a decade. Hence the benefits from cutting your price to gain market share are small--you couldn't satisfy that extra demand anyway.

But information good supply is like none of these. Fixed costs are high *and* fixed costs are sunk *and* marginal costs are near zero *and* capacity constraints are absent. Hence in order for a firm to earn profits in an information good market it needs to sell something unique, or to sell something at a price that no other firm can match: either differentiate your product, achieve cost leadership, or disappear.

Achieving cost leadership--making sure that your costs are less than the costs of your rivals--in an information good market is both very simple and very difficult. It is very simple because there is not much you can do. In a more standard industry firms achieve lower costs by working hard to make their operations more efficient: "using supply chain management, workflow analysis, and other tools to cut costs of parts, assembly, and distribution" (Shapiro and Varian, p. 28). In an information good market, the way to achieve cost leadership is to sell more units. Period. Unit cost is, roughly, inversely proportional to volume because the biggest costs are the initial, fixed ones.

Thus as Shapiro and Varian see it all of the supply-side factors that usually allow competitors to share markets, and that tend to keep prices elevated far enough above marginal cost for firms to earn a decent return on investment, are absent. The image of the information good sector that emerges is of a great number of bottles on the beach. Some of the bottles contain only one scorpion--a firm that has managed to successfully differentiate its product and find a strategy that is hard to copy and allows it to charge a premium price to those consumers with a high willingness to pay. The other bottles contain two or more scorpions, only the biggest of which can survive.


C. Network Externalities

So how then does a business become the biggest scorpion in the bottle?

The answer that Shapiro and Varian lean most heavily upon is that information goods tend to have powerful demand-side network externalities (as well as powerful supply-side economies of scale). The more units you sell--whether hardware or software--the more likely is the average user to find someone to help him or her set up the machine or the program, or find a product that works well with it. Not only is a high-volume product cheaper for the firm to make, it is more valuable for the user to buy. Hence "growth is a strategic imperative" because only a large volume can make the product valuable to the consumer (Shapiro and Varian, p. 13).

If a firm can achieve such growth on its own, then it is in a wonderful competitive position. To be first to the market, to have a substantial installed base while your competitors are still vaporware, to have a network of formal and informal support systems, users' groups, and add-on products available is to be in an almost insurmountable competitive position. Look at how many arms Microsoft has had to twist over the past four years and how much it has had to invest in writing code and in advertising in order for Microsoft to draw even with Netscape in the web browser market. I do not believe that any other company could have done it, for no other company had both the control and the incentive to use all of its leverage to try to erase Netscape's initial advantage.

But almost no firms can establish such an initial growth advantage on their own. In order to persuade consumers that a firm's products will be dominant--in order to become the product that is "expected to become the standard" and so "will become the standard"--a firm needs partners and allies. Hence, as Shapiro and Varian put it, "firms must focus not only on their competitors but also on their collaborators. Forming alliances, cultivating partners, and ensuring compatibility (or lack of compatibility!) are critical business decisions" (p. 10).

Not only are information good-selling firms scorpions in bottles, they are scorpions that need to cooperate with each other if they are to grow large.

Microsoft executives today like to point out to regulators and other potential ill-wishers that no company has yet managed to dominate multiple consecutive edges of the computer industry technology frontier. IBM dominated the era of mainframes, but the minicomputer of choice was Digital Equipment's VAX. Apple dominated the first generation of personal computers, but IBM dominated the second. Now Intel processors and Microsoft software dominate the market, but even they have something to fear from the unknown future. Reading Shapiro and Varian's book and thinking about the problems of firm business strategy in the network economy makes you understand why Andrew Grove, CEO of Intel--one of the largest and most successful of the scorpions--titled his memoirs, Only the Paranoid Survive.

III. How the Network Economy Is Changing Everything

Kevin Kelly sees many of the same phenomena as do Carl Shapiro and Hal Varian. In most cases I prefer Shapiro and Varian's formulations--I am, after all, part of their discipline: a professional economics professor, not a professional visionary. I find it more useful to think of network externalities as generators of powerful demand-side increasing returns to scale, and less useful to think in Kelly's terms of:

Rule 3: Plentitude, Not Scarcity. As manufacturing techniques perfect the art of making copies plentiful, value is carried by abundance, rather than scarcity, inverting traditional business propositions (Kelly, p. 37).

And I find Shapiro and Varian's argument that:

content providers tend to be too conservative with respect to the management of their intellectual property. The history of the video industry is a good example. Hollywood was petrified by the advent of videotape recorders. The TV industry filed suits to prevent home copying of TV programs, and Disney attempted to distinguish video sales and rentals through licensing arrangements. All of these attempts failed. Ironically, Hollywood now makes more from video than from theater presentations for most productions. The video sales and rental market, once so feared, has become a giant revenue source for Hollywood. When managing intellectual property, your goal should be to choose the terms and conditions that maximize the value of your intellectual property, not the terms and conditions that maximize the protection (p. 4).

more understandable and more convincing than Kelly's:

Rule 4: Follow the Free. As resource scarcity gives way to abundance, generosity begets wealth. Following the free rehearses the inevitable fall of prices, and takes advantage of the only true scarcity: human attention (p. 47).

But the two sets of ideas are, at bottom, the same: Shapiro and Varian's use of the basic version of your product as a marketing and information tool is the same idea as Kelly's "following the free." Shapiro and Varian's focus on demand-side network externalities is the same idea as Kelly's "value is carried by abundance."

Moreover, your mileage may vary: just because I find Shapiro and Varian's style of argument and tone more congenial than Kelly's doesn't mean that you will agree. So anyone reading New Rules for the New Economy who starts to feel uneasy should be reassured by the fact that in a large number of chapters what are in outline and gross the same conclusions are reached in the more sober-toned and business strategy-oriented Information Rules. Specifically, Kelly's chapters 2, 3, 4, 5, 7, and 9--"Increasing Returns," "Plentitude, Not Scarcity," "Follow the Free," "Feed the Web First," "From Places to Spaces," and "Relationship Tech"--all cover the same phenomena (although from a very different rhetorical perspective) as the parallel sections in Shapiro and Varian.

But what about the rest of the book? The parts that do not have analogues in Shapiro and Varian?

The first way in which New Rules for the New Economy deviates markedly from Information Rules is that Kevin Kelly sees strong pressure for decentralization. Where Shapiro and Varian always adopt the standpoint of a corporate strategist trying to decide what to do, Kelly wonders who the strategist is. The conclusion that he reaches is that information--the power to know enough to figure out what is the right thing to do--is going to be much more accessible much earlier much closer to the fringes of the organization than in the past. Where management in the past may have consisted in telling people what to do, management in the future may consist much more in making sure that people at the periphery understand the goals of the organization and picking people who will be good strategists as more and more decisions move outward from the center (which was the only place where the big picture could be seen) to the edges (where both the big picture and the details in which hides the devil can be seen).

Now this point can be overdone. Appropriate decentralization was always a key task of managers, and a key characteristic of successful corporations. But Kelly believes--and I think that he is right--that it is about to become much more so.

The second extra point that Kelly sees is that successful organizations tend to become complacent and defensive organizations. Kelly sees a willingness to take many risks--up to and including cannibalizing very successful businesses--as a principal characteristic that successful organizations will have in the future. In some ways this point is a commonplace: not a history of IBM vs. Microsoft doesn't place a large part of IBM's loss of leadership in microcomputers to its executives' reluctance to allow its microcomputer divisions to compete with its older and highly successful lines of business.

Third and last, Kelly advises us all to "seek opportunity, not efficiency." The point is that more of enduring value is created by finding new things to do than by figuring how to do old things better. And here Kelly truly earns his pay as a visionary, leaping from topics like postmodernism to stock options to networks to Bangladesh's Grameen Bank to the efficiency of inefficient exploration to learning curves and compound interest to Steven Jay Gould's Great Asymmetry of life. I cannot assess the value of these ideas--and Kevin Kelly can't assess their value either, that's part of the point.

But I do see a number of things that make me believe in the need for vision. To provide just one example, consider Netscape's decision to release not just its product but the source code for Netscape Navigator to essentially everyone who is interested, in the hope that outsiders will test, modify, use, and improve the program. How will this earn profits for Netscape shareholders? The view appears to be that a better Netscape Navigator will make more people stop by Netscape's portal to the world wide web and more people buy Netscape's other--not free--products. With the public release of source code, Netscape's profit model from its browser appears to be turning into a combination of an old-fashioned community barn-raising and a public radio fund-raising drive. Public release of source code is a new model for writing programs, and is very definitely not business as usual for selling software.

What forces are at work that might be pushing the software industry in the direction of open-source Navigator as a web browser, and free LINUX as an operating system, and free APACHE as a web server? Here are people who are not pursuing what we think of as economic efficiency at all, but who are pursuing opportunities.

It is probably the case--but it is not yet totally clear to me--that technologies change, but that economic principles do not...

Professor of Economics J. Bradford DeLong, 601 Evans, #3880
University of California at Berkeley
Berkeley, CA 94720-3880
(510) 643-4027 phone (510) 642-6615 fax

This document:

Search This Website