The general theoretical proposition that the worker who loses his job in one industry will necessarily be able to find employment, possibly after appropriate retraining, in some other industry is as invalid as would be the assertion that horses that lost their job in transportation and agriculture can necessarily have been put to another economically productive use.

— Wassily Leontief
... today's tech-savvy, well-compensated worker could become an expensive anachronism as tomorrow's technological advances offer new opportunities for slashing costs and improving economies of scale. A world filled with smart computers, all linked via the Internet, could easily undermine whole sectors of today's vibrant service and information industries. In the next century, lawyers, accountants, and brokers could be the secretaries, bank tellers, and mainframe operators of the 1980s.

— Business Week (1998)

This paper was presented at the International Conference on the Social Impact of Information Technologies in St. Louis, Missouri, U.S.A., October 12-14, 1998. The conference was sponsored by the University of Missouri-St. Louis and Southwestern Bell. The author's conference attendance was sponsored by Bellevue Community College's Distance Education Program and Social Science Division.

The Brief Reign of the Knowledge Worker:

Information Technology and Technological Unemployment

by Kit Sims Taylor

Long heralded, the era of the knowledge worker has finally arrived. With ever-increasing computing power on our desks, on our laps — even in the palms of our hands — and with instant access to information via the proliferating interconnections of the Web,— we substitute bits for brawn, search engines for sweat, and trend analysis for tedium. Let the clerk, the machinist, the sheep-shearer, and even the lower rungs of middle management fear the computer — but for us it is a tool.

The term knowledge worker is used here much the way Robert Reich uses symbolic analyst — to include professionals, upper-middle managers and above, and others who create, modify, and synthesize knowledge. This would include about 20 percent of the U.S. labor force.
With this new era has come such a proliferation of terms that we hardly know what to call ourselves or our new world. Peter Drucker coined the term knowledge worker in 1959, but left the definition rather fuzzy. In 1991 Robert Reich put us into the narrower — but more carefully defined — category of symbolic analyst. With our fingers on our mice and our minds in cyberspace, do we toil in Drucker's knowledge society, Manuel Castells' network society, or Wired's new economy? It matters little what we decide to call ourselves – we are the masters of the information technology revolution and the creators of the information that seem to be the driving force behind a long boom. Secure in an economy which finds ever-new uses for our mental creations, we fail to ask the obvious question: What comes after knowledge work?

There is clearly a rosy scenario afoot: the knowledge sector – with much higher productivity than other sectors of the economy – will expand rapidly in terms of employment as well as output; as most workers become high-productivity knowledge workers our incomes will rise proportionately; best of all, this will be accomplished within the institutional structure of capitalism as we know it today.

Unfortunately, this rosy scenario is based on five premises and four of these are incorrect. The first premise – the correct one — is that work of low skill and low productivity will disappear.

The incorrect premises are:

  • That knowledge workers primarily create new knowledge.

  • That demand for knowledge work will grow in pace with the increased supply and rising productivity of knowledge workers.

  • That we can make meaningful assessments of future labor markets by projections from today's labor markets.

  • That productivity increases, by themselves, will lead to increases in income.

These premises about the information economy, and the implications that can be drawn from them, will be examined in this paper.

Is There a New Economy?

Paul Krugman spars with Wired's Kevin Kelly in "New Economy? What New Economy?" and lays out his dim prognosis for knowledge workers in "White Collars Turn Blue."

Nicholas Negroponte's "Bits and Atoms " is one of the must-read essays on the digital revolution.
There are some – mostly economists – who would deny that we have entered a new economic era at all or that there are any discontinuities with the economic patterns of the past. MIT economist Paul Krugman – who is making a new career out of debunking the new economy – sees a fruitful future for non-knowledge workers whose work requires an automation-defying interaction with the physical world. But there is more than one way to shear a sheep. The exact path by which knowledge replaces physical work – or by which bits replace atoms – often comes from an unexpected direction:

  • Sheep-shearing robots may remain technologically unfeasible for some time, but Australian researchers have come up with an injection of natural protein that causes sheep to shed their wool.

  • A new mathematical model for airline scheduling is able to include all connecting flights when assessing delays. It could reduce delays by thirty percent. In addition to saving delay-induced expenditures in fuel and overtime – estimated to be $3 billion a year in the United States alone – the new model will allow our existing stock of airport facilities and airliners to serve more passengers.

  • Many technological improvements in medicine increase labor costs. But a new noninvasive test for breast cancer which is based on the skin's electrical field might cut the need for biopsies in half.

  • A new computer-guided system for finding defects in textiles will not just replace human inspectors. The inspection step is a serious bottleneck which restricts fabric companies to machinery that operates at 10 to 20 meters per minute. The new system will work at speeds up to 120 meters per minute. And it is more accurate than human inspectors. There is less waste and defects can be eliminated more quickly.

In all of these cases labor is being replaced by knowledge. Sometimes the labor that is displaced is outside of the knowledge-worker realm: sheep-shearers; construction workers and aircraft-assembly labor that will not be needed as we learn to create more passenger-miles of service with our existing aircraft and facilities; and fabric inspectors. There is also an indirect displacement of labor through less wasted fuel in the air transport industry and less wasted fabric as electronic defect sensors lead to quicker error-correction in textile production. Sometimes there is a displacement of the labor of knowledge workers as well: fewer and shorter air travel delays means more passenger-miles per pilot; reducing the number of biopsies decreases the demand for doctors and medical technicians.

The pace of the present technological revolution is far more rapid than any of the fundamental shifts of the past. The shift of labor from agriculture to manufacturing took over 100 years. We are still undergoing the shift from manufacturing to services which started almost 50 years ago. And now we face an ever-accelerating shift from service and manufacturing labor based on the speed and stamina of the sheep-shearer, the tact of the sales clerk, or the resistance to boredom of the fabric inspector to a higher level of service sector employment based on the ability to acquire, synthesize and manipulate knowledge and data. Only an economist wedded to Alfred Marshall's dictum that "there are no sharp corners in economics" could fail to see the discontinuity.

Employment in the New Economy — Prevailing Views

Joseph Schumpeter, in lectures at Harvard in the 1930s, claimed depressions were "a good cold douche" for the economic system.
But what of those who do see the discontinuity? What do they counsel? There are, of course, the Luddites who, with Kirkpatrick Sales, would smash the computers. And there are the ultra-Schumpetarians who would welcome a lack of employment opportunity for the unknowledged as a good cold shower that will bring them back to reality and force them to get the knowledge that will make them employable once again.

Let us turn, however, to the more thoughtful observers of this monumental shift. If there is an employment problem in the new economy, they say, it is a supply of labor problem. It is not a lack of jobs, but a lack of workers with the skills that are needed and prized by the information economy. Business journalists Bob Davis and David Wessel, in their recent book Prosperity: The Coming Twenty-Year Boom and What It Means to You, find growing inequality and lack of job opportunities for the unknowledged to be an education problem. Moreover, it is an education problem that the United States is rapidly solving through our dynamic and flexible network of community colleges. One of their chapters is even subtitled "How community colleges will foster prosperity and equality."

Reich is quoted on the dust jacket of Prosperity...: "Davis and Wessel make a clear and compelling case that it's now possible for America to achieve a broadly shared prosperity. The jury is still out on whether we will choose to do so."
Former Secretary of Labor Robert Reich also finds inequality and limited job opportunities to be primarily an education problem. But he is less optimistic than Davis and Wessel. Reich sees evidence that the richest fifth of the U.S. population is "seceding" from the rest of the country. They send their children to private schools then they oppose adequate funding of public schools. They live in residential enclaves with private guards and oppose the taxes needed to support police departments. Giving every child with the intelligence to become a symbolic analyst the education that it takes to become one should be a national priority. But Reich is skeptical that it will.

Castells warns that regions, castes, or social classes that are left out of network society will become economic black holes without value to capitalism as either producers or consumers: "If you are outside the network ... you don't even exist."
Sociologist Manuel Castells presents us with a thorough analysis of how the information technology revolution is fueling a " fundamental process of restructuring of the capitalist system." Informationalism, according to Castells, is a new "mode of development" but it still operates within the capitalist mode of production. Castells examines employment trends and concludes that "information technology per se does not cause unemployment," although "those individuals and groups unable to acquire informational skills could be excluded from the labor force or downgraded as workers."

What Davis, Wessel and Reich assume, and what Castells arrives at from projection of present trends, is that there will be no shortage of job opportunities in the knowledge sector – that knowledge work will be available for all who have the intelligence and the education to perform it. Let us examine the meaning of such an assumption. For the number of knowledge-work jobs to remain the same, the demand for the output of knowledge workers must increase as fast as the productivity of knowledge workers. For the knowledge sector to absorb the retrained cast-offs of other labor markets and to absorb the high percentage of the young who will acquire the requisite knowledge and skills, the demand for the output of knowledge workers must increase much faster than the productivity of knowledge workers.

The Demand for Knowledge Workers

A few numerical examples will help us to understand what is involved in such a shift. If knowledge workers – or symbolic analysts – presently make up 20 percent of the labor force, what will it take for us to make up 30 percent of the labor force within 20 years? If the labor force grows by 1 percent per year (which is the approximate growth rate of the U.S. labor force over the last ten years), and there is no increase in the productivity of knowledge workers the demand for the goods and services produced by knowledge workers must increase by slightly over 3 percent per year. If we add the moderate assumption that the productivity of knowledge workers will grow by 2 percent per year, then the demand for their goods and services would need to grow by over 5 percent per year.

While our desire for the products of knowledge workers may well be infinite, desire is not the same as demand. Effective demand requires the ability to purchase the product as well as the desire to do so. Demand — unlike desire, want, or even need — is limited by income.

So where does the demand for the goods and services produced by knowledge workers come from? We don't go over to the knowledge section of the supermarket to pick up a pound of abstract knowledge. The knowledge that we purchase is either embodied in other goods and services — such as the engineering that went into a refrigerator — or comes in a specific form as a computer game or a video. Some products are more knowledge-intensive than others. Let us use as examples software and chicken.

To keep the distinction clear, we will assume that the software is distributed over the Internet with no packaging, paper manual, etc. It is a 'pure' knowledge product produced by programmers and sold by marketing managers – all knowledge workers – with no physical goods or work by unskilled workers involved. Chickens are close to the other end of the spectrum. They require corn and other feed, transportation of the corn, abundant unskilled labor in the chicken disassembly plant, transportation of the chickens to the supermarket, and so forth.

Sticking with this example, there are three ways in which our demand for the products of knowledge workers can increase. First, our income could increase allowing us to purchase more software and more chicken. Second, our tastes could change in such a way that we purchase more software but less chicken. Third, new knowledge could be substituted for unskilled labor and other economic resources that are used to produce chicken: agricultural research might lead to chickens that require less feed per pound of weight gain; robots might replace unskilled workers in poultry-packing plants; or poultry scientists may develop a drug that causes chickens to quickly shed their feathers. The third way may well turn out to be the most important source of growth of demand for the products of knowledge workers. It also would lead to a more rapid disappearance of unskilled jobs and other jobs not requiring knowledge workers.

We must also consider two factors that differentiate knowledge from other products. First, the use of knowledge is not exclusive. My use of a 'unit' of software does not preclude your use of it. Once the software is produced, any number of people can use it. Second, my use of software or other knowledge today does not preclude my use of it tomorrow. Knowledge lasts until it is replaced by better knowledge. This means that there are some other important differences between the demand for software and the demand for chicken. Even if we are content with the same type of chicken year after year, the suppliers will need to continue producing it. But if we continue to use the same software year after year, there will soon be little or no market for it. Thus an increasing demand for knowledge products really means an increasing demand for new knowledge products, not just the continuing application of existing knowledge. The same consideration applies to our indirect consumption of the knowledge embedded in chicken. If poultry scientists develop a new mixture of feed that decreases the amount of feed necessary to produce a pound of chicken, this research does not need to be repeated year after year. Once the knowledge exists it can potentially be used forever.

The Supply of Knowledge Labor

Now we can turn to the potential supply of knowledge products. This will depend on the number of adequate workers offering to work, the number of hours that each knowledge worker seeks to work, and the productivity of their labor. We can assume that the number of workers seeking employment as knowledge workers will increase due to both push and pull factors. The major push factor will be the disappearance of other types of employment. A major pull factor is the interesting and pleasant nature of much knowledge work compared to other jobs. And, unless there are significant increases in the wage rates for knowledge labor, it is not likely that the number of hours each one seeks to work will decrease.

Chart: Knowledge Work
– Supply, Demand, and
The relationship between the supply of knowledge labor and the supply of knowledge products is the productivity of knowledge workers, or how much the average knowledge worker produces in an hour. If productivity increases rapidly, it means that the same number of workers working the same number of hours can produce significantly more than they could before. Or, a smaller number of workers could produce the same amount that was produced before. The diversity and non-comparability of the goods and services produced by knowledge workers makes productivity difficult to measure with any accuracy – we can't reduce their output to some measurable standard like tons of steel produced per day – but the concepts of productivity and productivity growth are important nonetheless.

This brings us to the key question of this paper: Is it likely that the effective supply of knowledge work will grow faster than the demand for knowledge products? In the past this has occurred with both agriculture and manufacturing. As the productivity of farmers and farm workers grew faster than the demand for agricultural products, the number of farmers and farm workers had to fall. When the productivity of manufacturing workers grew faster than our demand for manufactured products, the number of manufacturing workers necessarily shrunk. But there are two important differences. Employment in agriculture fell as employment in manufacturing was growing; employment in manufacturing fell as employment in the service sector was growing. And in both agriculture and manufacturing the slow pace of change made it easier for the growing sector to absorb the labor that was being cast out of the shrinking sector. The pace of technological change is much faster now. And there is no apparent sector that can absorb the labor that the knowledge sector casts off or the labor cast off by other sectors that the knowledge sector fails to absorb. When we finally get around to asking "What comes after knowledge work?" we have to admit that there is no answer.

What do Knowledge Workers do?

As knowledge workers, we like to think that most of our work involves the creation of new knowledge — of knowledge that would not exist without our mental efforts. Unfortunately, this is actually a fairly small part of our work. When we examine the work pattern of knowledge workers, we find six more or less distinct types of work:

Chart: What Do Knowledge
Workers Do?
  1. Routine work that is hard to separate from knowledge work. Formatting an article, for example, is work that might be done by a typist, but would be done by the knowledge worker when that takes less time than preparing the document and formatting instructions for the typist.

  2. Networking, promoting, socializing.

  3. Finding the data needed to produce the knowledge.

  4. Creating what others have probably already created when this would take less time than to search, find, and appropriate what has been produced by others.

  5. Truly original knowledge work — creating what has not been created before.

  6. Communicating what has been produced or learned.
The most important distinction is the difference between '4' and '5.' We may spend much of our time creating knowledge that is new to us, but is the same or similar to 'products' that have been created by other knowledge workers. The fate of most knowledge workers is a worklife in which we are constantly reinventing the exam question, the flood insurance clause, the advertising copy for a sweater ad — producing goods and services that are new to us but not new to the greater society. Such work may well be pleasant and fulfilling. The knowledge worker is indeed engaged in the creation of knowledge. But the knowledge worker is only being paid for it because creating knowledge that exists elsewhere is presently cheaper than finding it.

Networked Productivity

There is a common assumption that – short of cosmic leaps in artificial intelligence – knowledge work cannot be automated. Knowledge work, of course, involves creativity and computers are just dumb beasts with very reliable memories. They can assist the creative process, but cannot take it over. That assumption is likely true. But it is no longer relevant. When the computer was a solitary machine, standing alone, one to the desk of each knowledge worker, the assumption was applicable. The Net, however, has changed everything. As the interconnections among computers increase in complexity and speed, our computers will offer us immediate access not only to our own past work, but to the work of all the other knowledge workers in the world.

Indeed, the interconnections of the network may provide much of what we have been seeking in expert systems software, and be cheaper and more reliable as well. One expert systems program used for bond trading includes more than 2,000 rules. The program attempts to reproduce the thinking patterns of a particularly proficient bond trader. But it will soon be feasible to compare the actions of thousands of bond traders and distill the common patterns that have been the most successful.

Finding the 'intelligence' in the network that we once hoped to develop as artificial intelligence and expert systems software does not depend on some yet unrealized technological breakthrough. It can come from the continuation of present trends in speed, interconnectivity, and search routines. In particular, there are three ongoing technological trends that we should note:
  • More bandwidth at lower cost: Many of us now have cable connections in our offices. Our homes are next, with DSL (Digital Subscriber Line) and ADSL-light (Asymmetric Digital Subscriber Line) being offered by the phone companies and even faster connections offered by our TV cable providers.

  • Breakdown of the server/client distinction: Now we make information accessible to others by uploading it to a Web server computer. In the near future we will simply leave it in accessible sections of the hard drives on our desktop computers. SETI is already undertaking an experiment in distributed computing that taps the unused capacity of over 100,000 computers.

  • Improved pattern recognition: Search engines become more sophisticated every month. It is also possible that new protocols, such as Extensible Markup Language (XML), will standardize the methods of getting more information about documents (meta-information) into the documents themselves.

While these developments will have a limited effect on truly original knowledge work, they will have a drastic effect on the knowledge products that we create but that have already been created by others. At present it is easier to write that contract clause, exam question, insurance policy clause, etc. ourselves than to find something close enough to what we want from elsewhere. But, based on the continuation of present technological trends, it will soon take less time to find it than to create it. While most of us do not like to admit that much of our creative work involves reinventing the wheel an honest assessment of our work would indicate that we do far more 'recreating' than creating.

The Internet and other information technologies have already had a significant impact on knowledge work, even though some of these technologies are still in their infancy. In the near future we can expect information technology to have an even greater impact on all of the aspects of knowledge work that were identified above:
  1. Routine work that is hard to separate from knowledge work: The stand-alone computer has already had a major impact on routine work, but the employment shrinking effects have fallen more on clerical workers than on knowledge workers. This trend will continue as we network more of our computer activities. For example, rather than print out something that needs to be reproduced, then give it to an office worker who — in turn — fills out a form and sends it to the campus copy center, instructors at my college now fill out a form on screen and send documents directly to the copy center over our LAN. Perhaps we only eliminate two minutes of the instructor's time and four minutes of the clerk's time for each document, but it adds up to substantial labor reduction.

  2. Networking, promoting, socializing: Email, chat rooms and electronic white boards are presently eliminating a lot of travel and waiting from our networking. As this technology becomes easier to use, and as we become more comfortable with it, more face-to-face communication will be replaced with electronic contact. Many knowledge workers may not like it, but when the institutions that employ us are convinced that electronic connections will do the job, we will find our travel and conference budgets slashed.

  3. Finding the data needed to produce the knowledge: The Net has already had a major impact on this aspect of our work. When we consider that as recently as five years ago most of us hadn't yet begun to use the Internet, the potential for the next five years seems almost unlimited. The Internet is already beginning to replace the law library. It will increasingly be used for medical diagnoses, journalism, engineering, and practically every other professional activity.

  4. Creating what others have probably already created: This aspect of knowledge work will practically disappear. It will become a form of data gathering or knowledge mining rather than knowledge creation. Whirlpool now puts the engineering details of the many components used in its appliances on its Intranet. "In this way an engineer in Brazil who wants a specific aspect for a new refrigerator can look up the website and borrow an idea that has already been invented for a product in the US or Europe," one of their executives noted. Whirlpool expects to increase productivity of its product designers by 30 percent. For many knowledge workers, creating what others have already created is the major portion of our work. It is also the aspect of our work that will be most impacted by information technology trends.

  5. Truly original knowledge work — creating what has not been created before: Information technology products such as spreadsheets make it much easier – and faster – to test our concepts. Future gains here may be limited, but the unfortunate reality is that this is a small proportion of our work time.

  6. Communicating what has been produced or learned: The Internet is emerging as a universal communication medium. As voice and video become easier to integrate with text and graphics, the process of communicating knowledge will meld into the process of creating it.

Who Gets the Income?

In his pathbreaking new book, Created Unequal, James K. Galbraith shows how macroeconomic policies that permit unemployment to remain high have led to increasing inequality among wage and salary earners. For his final New York Times column, Peter Passell chose the topic of the distribution of income. "For the last quarter century," he wrote — referring to the U.S. economy — "virtually all the bounty of growth has gone to the educated, enterprising and already-affluent." We can identify two major forces — both moving in the same direction — that operated to skew the distribution of income. One was the increase in real interest rates, which benefitted the owners of money. Interest made up 9.2 percent of personal income in 1973 and 13.3 percent by 1995. The share of total income earned as wages, salaries and fringe benefits fell from 74.3 percent to 71.5 percent over the same period. The other was the decline in the value of routine labor relative to the value of college-educated labor. In 1979 the median male college graduate earned 42 percent more than his high school graduate counterpart; by 1998 his college education premium had risen to 89 percent.
ISDN phone are creating a superstar effect among voice-over actors. "Advances in technology have helped the most popular voice stars bite off a bigger chunk of the business." Another — less expected — factor in increasing earnings inequality has been growing inequality of earnings within occupations. This trend has been particularly pronounced in the white collar professions. One of the causes of this intra-occupation skewing is the proliferating computerization of business. According to Robert H. Frank and Philip J. Cook in The Winner-Take-All Society, the ability of computerized production processes to gather data as well as produce goods increases the leverage of the most able participants: "No matter what new organizational forms ultimately emerge, the cumulative effect of these changes will be to increase still further the leverage of the economy's most able performers."

Frank and Cook's observations were made in 1995, just as many knowledge workers were taking our first tentative clicks into the Internet. The technological trends noted earlier in this paper can only serve to accelerate these earnings trends. We will not only have superstar CEOs and superstar attorneys, but superstar fire-insurance-policy authors and superstar egg-marketing experts whose Web-distributed services will bring them greater earnings while curtailing opportunities for others in those fields.

Chart: Who Gains from
Productivity Growth?
Whenever productivity increases someone gains, but it is not always the worker. Labor market conditions — notably the unemployment rate — , the degree of competition among firms, and institutional factors such as the strength of government will all affect the distribution of the gains from increased productivity. With sufficient competition productivity gains may be spread throughout the society via lower prices. With low unemployment rates and/or strong unions they will be spread via higher wages. Government might capture the gains through taxation and spread them to society via spending on education or health care. If unemployment is high, unions are weak, firms face little competition, and government lacks the ability to increase taxes, productivity gains will accrue to the firms' managers as higher salaries and to the firms' owners as higher profits.

The complex linkages between productivity and wages are explored in "The Anatomy of Capitalism" — a chapter of my online textbook.
If the high productivity growth stemming from the new information technologies leads to even moderate technological unemployment, it is likely that the gains will be captured by a few superstars, a handful of top managers, and the owners of the firms. The same network effects that make a product more valuable to us when others use it as well (such as email systems and computer operating systems) have the effect of reducing competition. And the very technologies that make distance irrelevant give many firms a global choice of locations which weakens the power of government.

In September (1998) a fall in demand in Asia led Motorola to stop construction of a half-built semiconductor plant that was to have cost $3 billion.
An increasing concentration of income and wealth is not only an unfortunate development in terms of equity, but will also tend to slow economic growth. People of average income spend most of it on consumer goods; people with very high incomes will invest a large part of it. If consumption growth slows, there will be fewer investment opportunities. If our average consumer cannot afford gobs of the new products that include microchips, Motorola, Intel and IBM are not going to invest in new production facilities.

Shock Absorbers

If we were to look at nothing other than present economic statistics for the United States, we would be left with a quite different picture than the one that I have presented. One may well ask why the current profusion of new technology has not yet led to massive unemployment. After all, most of today's knowledge workers are equipped with computers. Most of us are already wired to the Internet. Productivity growth has been respectable lately, but not astounding. Unemployment has been remarkably low, with rates we have not seen since the 1960s. Toss in low inflation and it is easy to see where rosy scenarios come from.

"Evaluating Knowledge Worker Productivity," a report prepared for the U.S. Army, explores the complexity of productivity measurement.
One difficulty with trying to forecast the future economy from today's economic statistics is that our statistics are themselves rooted in the old industrial economy and are not adequate for a knowledge-based economy. We measure productivity in any industry by dividing the output of the industry by the number of hours worked. This works fairly well for steel and concrete as we have unambiguous measures of output. But when we try to measure productivity in the financial services industry, the telecommunications industry, or in government services such as education or public health there is no physical measure of output.

However, even if we could measure productivity growth adequately we would probably find it lower than a quick perusal of technological advances would lead us to predict. Productivity gains from new technologies can only be piecemeal until we adapt our production practices to them. While advocates of rapid technological change decry our attachment to yesterday's production practices, it might be more fruitful to view the various dimensions of this resistance to change as shock absorbers that help limit the pace of technological change to a rate more consistent with the abilities of society to tolerate change. Our current shock absorbers include inertia, the artisan attitude, a vast analog backlog, and a serious property rights puzzle.

  • Inertia: New technologies are at first superimposed on existing production systems. Don Tapscott calls this "paving the cow path." The word processor was first used as a faster and fancier typewriter: early users failed to appreciate word processors as part of a digital data and text storage system. Sometimes the inertia comes from the customer. It took some bank customers quite a while to realize that the ATM was often more convenient than visiting a bank teller.

    As more firms realize the cost savings from adapting production processes to take advantage of information technology, inertia will be harder to sustain. Aetna is attempting to get more medical insurance claims submitted electronically. Claim processing costs can be cut by 90 percent with higher degrees of accuracy. The company claims that there is a 40 percent error rate in manually-processed claim forms but only a 1 percent error rate with electronically-processed forms. The inertia comes from the doctors, with over 60 percent still filing claims on paper. Aetna hopes to increase electronic filing by promising faster payment to doctors who use their automated system. Perhaps more banks will emulate this tactic and offer us lower fees when we use ATMs.

Thorstein Veblen's essay "The Instinct of Workmanship and the Irksomeness of Labor" was written 100 years ago but is still timely.
  • The Artisan Attitude: When workers who have some control over their pace of work — this would include practically all knowledge workers as well as many clerical workers, crafts persons, and others — get a new tool, our initial impulse is to use it to produce a better product, not to increase productivity in the sense of doing the same thing faster. The secretary would rather format letters more attractively — using bold face, italics, centering, and tables — than just produce more letters per hour. Workers, including knowledge workers, exercise what Thorstein Veblen called "the instinct of workmanship" whenever possible.

The potential impact of information technology on higher education has been explored in First Monday in articles by David Noble (February 1998), Michael Margolis (May 1998) and myself (September 1998).
    However, the knowledge worker's artisan attitude is in direct conflict with the employer's drive to cut costs. We can expect a long struggle as many doctors, journalists, professors, attorneys and other professionals try to integrate information technology with our concept of professional responsibility while our employers and/or competitors are using the new technologies to cut costs.

  • Analog Backlog: The language of the computer and the Internet is digital. Prior to the information technology revolution business data was analog. And, until the transition to a digital format progresses further, we will continue to produce some business knowledge in an analog format. Even some of the initial microchip-based technologies — voice-mail and the fax, for example — kept data in an analog form. Until these technological dead ends fade from the picture, business will need to employ large staffs of analog-to-digital translators. Even when we force new data into a digital format, there are still those file cabinets full of paper awaiting translation. Unfortunately, our current labor statistics cannot reveal how many jobs primarily involve analog-to-digital or digital-to-analog translation tasks. It is likely that this is a major part of the impetus behind the current employment boom.

    The analog backlog will eventually disappear. As businesses realize how much cheaper it is to get and keep information in a digital format, they will structure their interfaces in such a way as to limit their intake of analog data. is already providing an example of an almost completely digital business model. Newly produced business information will be increasingly digital. We can also expect rapid progress in automating the translation of information, including the paper in all those file cabinets, from analog to digital. For the customer who refuses to order by email or via a form on the company's Web page there are increasingly reliable voice recognition programs that can take phone orders with little or no human intervention. Computer vision is also becoming cheaper and more reliable. Look for New York Times feature articles on "The Last Receptionist" and "The Last Bank Teller" sometime between 2003 and 2008.

    As the increasing proportion of business data available in digital form eliminates many routine jobs, it will also increase the productivity of higher level knowledge workers. Today, knowledge workers spend a lot of work time blending information that comes in digital form with information that comes in analog form. When practically all relevant information is available in digital form, much of the work of such employees can be eliminated. Using the financial sector as an example, bank tellers will disappear. All of the information that they now process manually will be available immediately in digital form. The loan officers and financial advisors who use this information will not completely disappear, but large parts of their jobs — the parts that involve the blending of digital and analog information — will be automated and we will need far fewer of them.

  • A Property Rights Puzzle: One of the slogans of the Web zealots is "information wants to be free." Yet it is a basic tenet of capitalism that every productive resource have an owner. When information was embodied in a physical form such as a book or a video cassette our industrial concept of property rights sufficed. Now that information has freed itself from a physical form and become disembodied bits that can be transported and reproduced almost costlessly, a lot of effort and expense goes into database firewalls, password protection systems, and other methods of limiting access to information.

    As more firms learn that the information in their computers has value outside of the firm, and as micropayment systems become cheaper and easier to use, the effect of this particular shock absorber will diminish. It will become simple for an insurance firm to offer other insurers the right to use a particular paragraph from one of their policies for a payment that may only be one or two pennies per policy. The purchaser will find locating and purchasing the paragraph cheaper than paying one of their knowledge workers to spend 15 minutes writing a paragraph that accomplishes the same thing.
These shock absorbers will slow — but not stop — the automation of information transmission. Some sectors will be more difficult to effectively digitalize than others. Several firms are currently attempting to automate medical records. They are encountering elements of all of the shock absorbers. Current data is either in non-digital form — including handwritten records — or in digital forms that predate the Internet. Much of the data is owned by various insurance firms who are resistant to releasing it to other firms. And many doctors are resistant to change: one orders doctors around. Health care isn't like banking, airline ticketing or other industries, in which hourly workers must accept whatever new technology lands on their desks. If physicians don't trust a new computer system, they will attack it with all the vengeance of white blood cells fighting an infection.
But these barriers are outweighed by the potential returns. Medical paper-work costs $200 billion a year, or over 2 percent of the U.S. GDP. If Healtheon and the other firms pursuing this business fail, other firms will emerge to chase such high rewards.

The Institutional Challenge

Robert Heilbroner elaborates on the distinction between activities and work in his essay "The World of Work." Unfortunately, this essay is not available online.
Our socioeconomic system faces both an opportunity and a challenge. The opportunity is the possibility of an abundant flow of goods and services with very little work — in the traditional sense — on our part. The truly creative activities that cannot be automated will be challenging and fulfilling — cleaning up the environmental damage from the industrial era; finding ways to extend life and health; space exploration; writing operas; etc.

Examining the types of institutional changes that could meet the needs of the network economy is beyond the scope of this paper. Several potential solutions are explored in Sally Lerner's 1996 paper, "The Future of Work in North America: Good Jobs, Bad Jobs, Beyond Jobs."
The challenge lies in redesigning our economic institutions to break the link between production and income. This link has served capitalism well for the last two to three centuries. We applied our labor, our land, and/or our capital to the production process. In return we earned wages, rent, or profits. These earnings supported our demand for goods and services. Our demand, in turn, allowed the production process to continue. As we applied more capital and knowledge to the production process, the productivity of our labor increased. When firms were small, competition among them assured that increasing productivity led to lower prices. When this occurred with basic goods such as food, clothing or transportation, the lower prices meant an improved standard of living for most members of society. When unemployment was low, the competition for workers assured that increasing productivity supported higher wages, again leading to an improved standard of living for many.

The rising real wages — whether they came from falling prices or an increase in money wages — fed demand. With demand increasing, higher productivity was used to boost output, not to reduce overall employment. And rising wages plus growing demand also pushed firms into finding more ways to increase productivity — usually by introducing new equipment a rapid rate. We enjoyed the effects of a virtuous circle of growth in productivity, wages, and demand.

Growth was not always smooth. There were periods in which the linkages from production to income to demand and back to production failed us. When the failure was serious enough and long enough it led to major institutional changes. During the last quarter of the 19th century decreases in the prices of basic goods and services such as steel and shipping were sufficiently steep to force waves of bankruptcies and bring new investment to a halt. Oligopoly and antitrust laws were among the new institutions that were subsequently created — allowing firms once again to expand with some hope of profits but with legal constraints on their power.

A more severe breakdown of the basic linkages of capitalism occurred in the 1930s. During the 1920s productivity in manufacturing had increased rapidly, but manufacturing wages had not. Out of the Great Depression came a much larger government role in influencing the distribution of income. In the United States this included laws facilitating the formation of industrial unions, social security, a minimum wage, minimum agricultural prices, and numerous other interventions. John Maynard Keynes provided a theoretical justification for this new form of capitalism in which government was responsible for maintaining sufficient levels of income and demand to keep unemployment low and growth rates high.

Unfortunately, a larger government role in the distribution of income is a rather unpopular idea at present. Even in countries whose citizenry expect and endorse a major government role in income distribution — France and the Scandinavian countries, for example — the ability of firms to leave and take jobs with them has limited the effective scope of government action. Yet the leaps in productivity inherent in the new information technologies cannot be realized without markets for the oceans of knowledge-based goods and services that we will soon be capable of producing. If we leave the distribution of income primarily to market institutions, the demand for knowledge products cannot keep up with the effective supply of knowledge products. It would be far better to begin a dialogue now on the economic institutions that we will need in the near future rather than wait — as we have done in the past — for a depression or other economic calamity to force institutional change upon us. The last depression brought us the New Deal programs and Keynesian economics that helped fuel consumer demand in the post-WWII economy; but it also brought us the deadly detour of fascism.

What comes after knowledge work? The choice is ours. If we fail to adapt our economic institutions to accommodate the extremely high productivity that will become possible, we may well face constant depression combined with a distribution of income that will be more extreme than we find today in countries such as Brazil and Mexico. Or we can enjoy lives without toil; lives in which we can take an abundance of goods and services for granted; lives without work as we know it today; lives filled with creative and fulfilling activities.

Kit Sims Taylor teaches economics – mostly online – at Bellevue Community College in Bellevue, Washington. Mr. Taylor's current projects include the development of a new course on The Economics of Emerging Technologies which will first be offered (online) in the Winter Quarter of 1999. Email address:
Photo by Heather Marks

Sources (In Order of Appearance)

Leontief Quotation: This turned up on the Internet when I was searching under "technological unemployment." I have not been able to locate the exact source.

Business Week Quotation: Christopher Farrell, Ann Therese Palmer, Seanna Browder, "A Rising Tide," August 31, 1998.

Kevin Kelly’s Interview with Paul Krugman: "New Economy? What New Economy?," Wired 6.05, May 1998.

Paul Krugman's "White Collars Turn Blue:" New York Times Magazine, September 29, 1996.

Nicholas Negroponte's "Bits and Atoms:" Being Digital , 1995, Knopf, New York.

Self-Shearing Sheep: "Sheep Made to Shed Fleece," New York Times, August 25, 1998.

Airline Scheduling Model and Noninvasive Cancer Test: "Developments to Watch," Business Week, August 10, 1998.

Textile Inspection Computer: Peter Marsh, "Quality control: Fault finding picks up pace," Financial Times March 20, 1998.

Schumpeter on Depressions: Robert L. Heilbroner, The Worldly Philosophers, 1953, Simon & Schuster, New York. Pg. 302.

Bob Davis and David Wessel: Prosperity: The Coming Twenty-Year Boom and What It Means to You, 1998, Random House/Times Business.

Robert B. Reich: The Work of Nations: Preparing Ourselves for 21st Century Capitalism. 1991, Knopf, New York.

Manuel Castells: The Rise of the Network Society, 1996, Oxford, Blackwell Publishers. See Chapter 4, "The transformation of work and employment: networkers, jobless, and flextimers." Interview with Manuel Castells ("Dark Side of the Boom") by Jay Oglivy in Wired 6.11 (November 1998).

Expert Systems Bond-Trading Software: Clive Davidson, Christine Downton’s Brain," Wired, December 1996.

More Bandwidth at Lower Cost: Michelle V. Rafter, "Wiring the Suburbs — One Cable Box At a Time," The Industry Standard, August 24, 1998.

SETI's Distributed Computing Experiment: Jeffrey Davis, "The Power is Out There," Business 2.0, Premiere (August 98).

Improved Pattern Recognition: Eric Hellweg, "XML: Cleaning Up Search Clutter," Business 2.0, Premier (August 98).

Internet Replacing Law Library: Geanne Roseberg, "Law Firms Turn to Technology to Manage the Information Explosion," New York Times, March 30, 1998.

Whirlpool Designers: Peter Marsh and Nikki Tait, "White goods: Whirlpool’s global clean-up," Financial Times, March 24, 1998.

Peter Passel on Income Distribution: "Economic Scene: Rich nation, poor nation. Is anyone even looking for a cure?" New York Times, August 13, 1998.

James K. Galbraith on Wage Inequality: Created Unequal: The Crisis in American Pay, The Free Press, New York, 1998.

Wages Decline as Share of Total Income: Lawrence Mishel, Jared Bernstein and John Schmitt, The State of Working America, 1996-97. 1997, M.E. Sharpe, New York. Pg. 66.

College Degree Premium Rising: Christopher Farrell, Ann Therese Palmer, Seanna Browder, "A Rising Tide," Business Week, August 31, 1998.

Voice-Over Superstars: Carol Marie Cropper, "The People Behind the Voice Behind the Products," New York Times, October 22, 1998.

Winner-Take-All: Robert H. Frank and Philip J. Cook, The Winner-Take-All Society. 1995, Free Press, New York. Pg. 55.

Motorola Suspends Investment in Chip Plant: Louis Uchitelle, "Capital Spending Takes a Break," New York Times, September 18, 1998.

Paving Cow Paths: Don Tapscott, Digital Economy, 1996, McGraw-Hill, New York. Pg. 82.

Aetna and Electronic Filing: Ron Winslow, "Aetna Launches E-Pay to Speed Payment of Claims to Doctors" Wall Street Journal, June 15, 1998.

Computer Speech Recognition: John Markoff, "Operator? Give Me the World Wide Web and Make It Snappy," New York Times, October 6, 1998.

Computer Vision: Geoffrey Smith, "Silicon Eyes: How chips that see will transform the way we live," Business Week, October 12, 1998.

Doctors Resist Computerization of Medical Records: George Anders, "Resistant Strain: Healtheon Struggles in Efforts to Remedy Doctor's Paper Plague," Wall Street Journal, October 2, 1998.

The World of Work: In Robert Heilbroner, Behind the Veil of Economics, 1988, W. W. Norton, New York.