Arnold Kling     Essays | Short Book Reviews | Favorite Links | Under the Radar | Home

A Meditation on Innovation

Arguing in My Spare Time, No. 4.20

by Arnold Kling

June 20, 2001


Innovation seems to be the most important force in the economy today. This may not have been true in the past. Even if innovation has always been important, this probably is more obvious today than was the case earlier in history.

The pace of innovation has accelerated. Presumably, this is because the perceived benefits of innovation have increased and/or its perceived costs have declined. Perhaps this trend will continue, as argued by Ray Kurzweil in The Age of Spiritual Machines. But one should examine that assumption.

This essay looks at innovation from a historical perspective. Then it offers some hypotheses about what might determine the rate of innovation in the future.

The Media Time Machine

To get some historical perspective on innovation, we might want to put some modern media figures into a time machine and send them back through history. Here are some examples of what I think this exercise would produce.

  1. Technology Review, January 1000 AD edition.

    MIT's Technology Review strives to keep its readers abreast of new technologies and their progress from the laboratory to the point where they change our lives. Imagine this publication in the year 1000 AD.

    When we asked a series of experts about the Next Big Thing, several of them mentioned crop rotation. Crop rotation continues to fascinate cutting-edge feudal lords, and we might see some new patterns being attempted over the next several hundred years.

    In addition, it looks like over the next five hundred years the design of the plow is going to undergo some really exciting incremental change.

    Finally, engineers are working on a new horse collar which they hope to perfect over the next three hundred years. This could allow horses to replace oxen in pulling plows. Although pundits predict a gradual adoption curve, lasting several centuries, this promises to be quite a breakthrough!

  2. Fast Company profiles Heinz Guderian, July, 1940.

    Fast Company celebrates the ordinary middle manager as hero. It profiles executives who use trendy management strategies to cope with a complex environment. Here is how this magazine might have written in the aftermath of the German Blitzkrieg's triumph in France in 1940.

    Not everybody feels warmly about Nazi Germany. For example, Winston Churchill, in his "finest hour" speech last month, complained about "the abyss of a new Dark Age made more sinister and more protracted by the lights of a perverted science."

    But there can be no denying that Germany is up to date when it comes to waging war. For a perspective on this remarkable fighting force, we interviewed Heinz Guderian, one of the men who helps put the Blitz in Blitzkrieg.

    As a Wermacht Panzer commander, Guderian says that he is a change agent who operates in an environment that encourages risk-taking and accountability. In the recent campaign to conquer France, his troops showed that what it takes to win in the game of tank warfare is flexibility, adaptability, and strong communication skills.

    To his men, Heinz is a sympathetic and caring leader who gives them a sense of empowerment. At the same time, he has the political savvy to impress the Fuhrer, who has been known at times to come across as a demanding boss.

  3. Family Feud goes back to 1950.

    This is a game show in which the contestants are given a survey question and are asked to come up with the most popular answers. It is not unlike Keynes' favorite metaphor of newspaper beauty pageant contests.

    Hi, Richard Dawson here. I'm your host for Family Feud. It's 1950, and it's time to play the feud!

    We surveyed a hundred contemporary economists and asked them this question. What single indicator would you use today to measure the level of economic development and production capacity of a country?

    (Contestant). Steel production! (Applause)

    (Dawson). Show me--Steel production! . . . Number one answer! Do you want to play or pass?

The point of juxtaposing Technology Review with the Middle Ages is to emphasize the dramatic change in the pace of innovation. For most of human history, innovation has been so sluggish as to be scarcely perceptible within a lifetime.

Today, we take new discoveries and new inventions for granted. Two years ago, I asked over a hundred high school students whether they expected another innovation to come along that, like the Internet, would sweep through the younger generation before older people could understand it. They were nearly unanimous in answering in the affirmative.

Hundreds of years ago, innovations diffused much more slowly. It could take generations for a new farming technique to come into widespread usage. It was difficult for a given generation, much less a given individual, to capture the benefits of innovation.

The point of juxtaposing Fast Company with the Wermacht was to emphasize the challenge of trying to look at history through modern lenses. In fact, in a book entitled Bureaucracy, edited by James Q. Wilson, the German Panzer army was held up as an excellent case of enlightened, decentralized management. To read the essay in that book, one would think that the German victory owed less to tanks and dive-bombers than it did to a modern, humanistic approach to management. We're talking about the Nazis.

At the time, most analysts put very little weight on the Germans' management system. They attributed France's defeat to such factors as the quality of German armor and the defeatism and divisiveness at the top of the French government, with military leaders especially reluctant to fight.

In his six-volume account of the second World War, Winston Churchill never discussed management training issues. As Prime Minister, he tracked progress by measuring the loss rate and replacement rate for equipment. He closely watched data on tonnage of merchant ships built and sunk in what he called the Battle of the Atlantic. He compared the rate at which enemy U-Boats were launched with the rate at which they were sunk. And he compared the rates of losses of airplanes with the rates at which they were manufactured.

The point of the Family Feud example is to emphasize that once upon a time there really was capital embedded in the capitalist system. The ability to deploy capital was symbolized by steel production. Indeed, most of the equipment that Churchill assiduously tallied was composed of steel.

For over one hundred years after Karl Marx, economists viewed the accumulation of capital as the essence of economic growth. When Robert Solow wrote his famous paper in 1956 on the causes of economic growth, change in the stock of capital was the central focus. Innovation made its appearance as part of the famous "Solow residual."

Capital or Innovation?

I do not think it is possible to understand the thinking of economists in the quarter century after the second World War without realizing that they looked at growth as determined by capital accumulation. The alternative view, that growth is determined by the ability of a society to encourage and deploy innovation, is very much a product of the last two or three decades.

The capital-centric view fell apart primarily because of what happened to the former Soviet Union. The Soviets appeared to be able to hold their own in capital accumulation. Steel production and related indicators would have told you that the Soviet Union was an economic dynamo.

However, the Soviet Union and its successor states now seem economically backward. It turns out that accumulating capital does not create a first-rate economic power if the system lacks the means to encourage and deploy innovation.

Another factor leading contemporary economists to look at innovation rather than capital is the advent of the information age. The revolution in information and communication technology has yielded goods and services whose value appears to owe relatively little to the physical materials used to produce them. Software provides much of the value in computers. The value of the information that passes through the Internet is much greater than the cost of the wires and switches that are its tangible components.

Today, we are becoming so convinced that innovation drives growth that we might be tempted to engage in a revisionist history that greatly reduces the role of capital in the Industrial Revolution. I worry about this.

In the case of the fall of France, I think that one should be cautious about tossing out the analysis of Churchill and others who saw the event first hand. Similarly, it might be unwise to view economic growth during the Industrial Revolution as consisting primarily of innovation rather than capital accumulation. Such a case may be plausible, but we should not dismiss the fact that economists who wrote during that period saw capital playing a central role.

One difference between a capital-centric view of growth and an innovation centric view is that the capital-centric view emphasizes delayed gratification. It took decades to build a railroad, a steel company, or other major businesses of the 19th century. The automobile industry was not built on Internet Time.

Three Stages of Growth

As Rosenberg and Birdzell emphasize in How the West Grew Rich, growth represents change. In order to enjoy economic growth, a society must overcome the resistance to change.

I am tempted to posit three stages of growth, with the fundamental obstacle to change differing in the three stages. The scheme can be laid out in the following table.

Stage of GrowthFundamental Obstacle to Change
Pre-industrialAbsence of Apparent Improvement
IndustrialNeed for Postponed Gratification
Post-industrialBasic Resistance to Change

Perhaps the best way to think of this table is to note that in the pre-industrial economy all of the obstacles to change are present. A new era emerges when one of the obstacles is removed. Therefore, it may be easiest to explain the table by working backward.

  1. Basic Resistance to Change

    People resist change. Try selling something to an organization. As an outsider, try selling a new product or service. Or as an insider, try selling a new idea. You will meet with resistance. Management guru Peter Drucker often is quoted as saying that a new product or process must provide an order of magnitude improvement (e.g., reduce cost by a factor of 10) if you want to achieve adoption.

    Even if change is good for an organization or community as a whole, powerful interests within that organization or community can be adversely affected. Miller and Heiman, in their book Strategic Selling, draw a sharp distinction between a result and a win.

    In order to close a sale, you need to demonstrate a win as well as a result. If a powerful interest is in a position to block the change, that change will not be implemented, even though it might be of benefit to the larger organization or community.

    In the literature on business and management, it is now conventional wisdom that successful incumbents rarely participate in change. For example, Joel Barker, author of Paradigms : The Business of Discovering the Future, likes to describe how traditional watchmakers rejected digital watches, traditional photo companies rejected photocopiers, and so forth. As a result, firms from outside the industry had to bring these innovations to market.

    Resistance to change is apparent in many cultural institutions and practices. Rosenberg and Birdzell, cited above, emphasize these factors in explaining why Western Europe overtook China and the Islamic world during the period from 1400-1850, in spite of the latter's advantages in scientific understanding and technological development at the outset of the period.

    Rosenberg and Birdzell argue that the central religious and political authorities, who tended to resist change, maintained strong control in China and the Islamic empire. In contrast, European mercantile city-states gained considerable autonomy, leading to competition and experimentation.

    Jared Diamond, in Guns, Germs, and Steel, also sees strong central control as a drawback for Chinese development. He writes, "a decision by one despot could and repeatedly did halt innovation." (p. 416)

    A compelling case could probably be made that the extremes of wealth that we observe today within the United States and more dramatically across countries can be explained by differences in receptiveness to change. People and countries that are willing and able to adapt to change most readily enjoy the highest standards of living. People and countries with the greatest resistance to change suffer the most poverty.

    A case along these lines appears to be made in Barriers to Riches, by Parente and Prescott. Although I have not read the book, its description suggests that it emphasizes resistance to change in explaining international differences in living standards.

  2. Need for Postponed Gratification

    Basic resistance to change is a powerful force in its own right. In an industrial economy, where change is embodied in physical capital that requires heavy investment, there is an additional difficulty that must be overcome. People need to have the means and the discipline to set aside current consumption in order to accumulate capital that will be productive in the future.

    Economists and others who lived in the latter part of the Industrial Revolution saw postponed gratification as important. They differed in their moral interpretation of the process of capital accumulation. Supporters of the capitalist system could picture it as a morality play in which people who saved in the short run were rewarded in the end. The Millionaire Next Door, by Stanley and Danko, revives this morality play for the present day.

    Keynes, never a big fan of the desire to save, saw the process of individual accumulation of capital through deferred gratification as a necessity during a period of rapid economic growth. But he mocked the notion that "always jam tomorrow, never jam today" will inevitably be the most socially moral outlook.

    Marx, of course, went well beyond mockery. He saw capital accumulation as exploitation, with the gratification of the workers postponed in order to enable the capitalists to consume more now and in the future.

    Rosenberg and Birdzell would agree with Marx that capitalists did not appear to suffer from squalor while they were accumulating their fortunes. However, they think that workers at least held their own. They argue that the Industrial Revolution should be viewed as a period in which economic growth reflected the gains from adaptability, and they de-emphasize deferred gratification.

    As I indicated earlier, I am not quite ready to join Rosenberg and Birdzell in throwing overboard the importance of deferred gratification. In trying to account for the difference between, say, Japan, and Latin American in terms of economic growth from 1860 to 1980, one would have to be struck by the high rates of individual saving in Japan as a key factor.

    Further evidence of the importance of deferred gratification is the fact that many of the most important institutions that developed in the industrial era were designed to mobilize savings. Modern corporations, banks, and financial institutions in general serve to encourage and channel saving.

  3. Absence of Apparent Improvement

    In the story of deferred gratification in the industrial economy, at least you gain in future consumption when you defer consumption in the present. You can see that the change you are undertaking today by building a railroad or installing a steam turbine will yield benefits to you in the future.

    Before the industrial era allowed change to be embodied in physical capital, the benefits of change were so small and incremental as to be imperceptible. The gain in productivity from re-shaping a plow is so small that one can understand why people would neither be motivated to attempt innovation nor feel compelled to spread news of incremental improvements.

    In the United States today, people view any problem as a reason to try something new as a solution. The processes by which goods and services are designed, produced, and delivered are extremely complex. The opportunities to make improvements by altering those processes are significant.

    However, if you took the state of knowledge backward 1000 years, you would not necessarily think that attempts to improve your living standard by trial-and-error innovation would be fruitful. When economic activity consisted of growing a limited number of crops with a limited number of tools, the scope for change also was narrow. If we were to put a group of business process re-engineering consultants into our time machine, chances are slim that they could provide any useful service to a feudal landlord. (Let's assume for now that they do provide a useful service to a modern CEO.)

    In a setting where innovations do not result in perceptible economic gains, the ability to defer gratification does not achieve very much. By setting aside current consumption, you get the occasional striking monument, such as the Great Pyramids or the Gothic Cathedrals. However, you do not accumulate capital. In fact, pre-Industrial architectural monuments are remarkable in part because they were so exceptional. It seems irrational to work on an edifice that will not be completed for over a century, and indeed over the course of several thousand years this did not happen very often.

Overall, I am suggesting that in the pre-industrial era people resisted change because change brought little benefit. In the industrial era, the benefits were evident, but they required people to save to accumulate capital, which in turn requires a willingness to postpone gratification. In the information age, innovation can yield benefit with less foregone consumption, because physical capital is less important.

A Spiral of Accelerating Change?

I have just said that in the pre-industrial era the invisibility of benefit was an impediment to change. In the industrial era, the long time that it takes to accumulate capital was an impediment to change. To the extent that these impediments were important in the past, they are less important today.

So, now that we're in the post-industrial era, and the only obstacle to change is the fear of change itself, are we going to see ever-accelerating change and economic growth? Ray Kurzweil, in The Age of Spiritual Machines, makes exactly that forecast. As I explained in an earlier essay, Kurzweil sees a long-term trend of exponential growth in information processing capability and he extrapolates this trend forward.

The question I raised in the earlier essay still troubles me. Can we continue to accelerate the pace of change? I see two challenges.

  1. The Filtering Process.

    Innovation is an error-prone process. For every success, there are many failures. We can think of these failures as the difference between "gross innovation" and "net innovation."

    Going forward, if we are going to increase the amount of net innovation, we either have to increase the amount of gross innovation and/or we have to develop filtering processes that produce a higher yield of net to gross. I find it almost inconceivable that we can achieve everything with more gross innovation. I suspect that improvements to the filtering process are absolutely essential.

    The Internet Bubble is a dramatic illustration of a filtering process gone awry. Large amounts of capital and talent were poured into flawed businesses.

    On the other hand, the correction was rather swift. One can argue that the rapid collapse of the bubble shows the ability of the private sector to correct its own mistakes.

    Overall, though, I believe that the Internet Bubble demonstrates the weakness of a process that depends on the intuition of venture capitalists and the hype of mass media. These processes will need to be replaced in order to improve the yield from innovations. My guess is that compared with venture capitalists' selection criteria, there are more reliable ways to select new businesses from the universe of proposals, based on statistical modeling. Once we start using statistical models to predict the success or failure of new businesses, my guess is that we will see some improvement in this area.

    My guess is that there are more reliable ways to obtain early feedback about the value of a new product or service than to create media-based trial balloons. At one time, favorable reviews in key trade publications could create a "hit" product in hardware or software. However, now that companies have manipulated this process, these magazines no longer have credibility with enough readers.

    I suspect that in the future, something like will take the place of trade magazines. However, a system of incentive to reward accurate reviewers probably needs to be established. Perhaps paid subscribers to sites will be given personal access to the best reviewers.

    In any event, I foresee more sophisticated and reliable methods for filtering innovations. Until these methods emerge, weaknesses in our ability to sift good innovations out of the universe of possibilities will constrain economic progress.

  2. Adoption challenges.

    Do you have Windows 2000(TM)? A web-enabled cell phone? A personal digital assistant? A digital camera? A DVD player?

    No one I know has all of these. Each of these technologies is used by only a small minority of my friends and associates.

    I myself use none of them.

    If these new products are going to have an impact on society, then people like me are going to have to get with the program. We need to adopt these innovations soon, so that the technology community can move on to create the next big thing.

Going forward, the pressures on our processes for filtering and adopting innovations are going to experience stress. For change to accelerate, there has to be continuous improvement in these processes. We need to get better and better at filtering and adopting innovations. Otherwise, these processes will become bottlenecks in the innovation system.

I think that the adoption and filtering bottlenecks will start to have a visible effect on the rate of technological progress within ten years. Perhaps we will look back in ten years and say that the deceleration of change had already begun at the time that I wrote my essays.

Another way of making the case that a deceleration is inevitable is to do so in terms of the rate of obsolescence. Innovations make old technologies obsolete. As the pace of innovation speeds up, the pace of obsolescence must speed up. How fast can it go?

Suppose that we date the modern record player from 1930. The CD player takes over 50 years later. Now, the DVD player is supposed to end the reign of the CD player after 25 years. Will the DVD be obsolete in 12-1/2 years?

In business computing, the COBOL programming language was popular for about 20 years. Visual Basic was a leader for 10 years. Does that mean that Java(TM) will be popular for only 5 years?

The "client-server" revolution in business computing was over so quickly that most companies missed it. That is, by the time they could start projects to implement client-server internally, the Internet made this model obsolete. The Internet became the medium over which companies implemented client-server computing.

In the pre-industrial era, technologies remained in place for centuries before going obsolete. In the industrial era, technologies lasted for decades. Today, we expect many technologies to be discarded in less than a decade. To keep the pace of progress accelerating, we have to develop the ability to discard important new technologies within the space of a year, or less. I just do not see how you get to that point.


Innovation appears to be the key to growth. In the world economy, people and systems that accomodate change tend to win out in the world economy.

It is plausible that the pace of innovation will continue to be high, and perhaps to accelerate a bit further. However, ultimately, the rate of innovation cannot continue to increase unless the processes for filtering, adopting, and discarding technologies can be improved significantly. Thus, although economic growth can continue, it is unlikely to accelerate exponentially, as it apparently has until now.