Learning by Doing: The Real Connection between Innovation, Wages, and Wealth

by James Bessen, Garamond Agency, Inc. (Creator)

An important study of the relationship between technology, skills, and economic inequality that answers some of the most pressing economic questions of our time

Today’s great paradox is that we feel the impact of technology everywhere—in our cars, our phones, the supermarket, the doctor’s office—but not in our paychecks. In the past, technological advancements dramatically increased wages, but for three decades now, the median wage has remained stagnant. Machines have taken over much of the work of humans, destroying old jobs while increasing profits for business owners. The threat of ever-widening economic inequality looms, but in Learning by Doing, James Bessen argues that increased inequality is not inevitable.
 
Workers can benefit by acquiring the knowledge and skills necessary to implement rapidly evolving technologies; unfortunately, this can take years, even decades. Technical knowledge is mostly unstandardized and difficult to acquire, learned through job experience rather than in the classroom. As Bessen explains, the right policies are necessary to provide strong incentives for learning on the job. Politically influential interests have moved policy in the wrong direction recently. Based on economic history as well as analysis of today’s labor markets, his book shows a way to restore broadly shared prosperity.

Genres: Economics, Business, Nonfiction, History, Technology, Education, Politics

312 pages, Hardcover

First published April 28, 2015

About the author

James Bessen, an economist and technologist, serves as Executive Director of the Technology & Policy Research Initiative at Boston University. He has also been a successful innovator and CEO of a software company. Bessen studies the major economic impacts of technology on society (see New York Times profile), writing academic papers, magazine articles, and books. His latest book, The New Goliaths (Yale 2022), argues that major firms’ investments in proprietary software systems have allowed them to increase their dominance of industries, slowing aggregate innovation and raising income inequality. Earlier work with Michael Meurer on patents identified the social costs of poorly defined property rights (see Patent Failure, Princeton 2008), including the first evidence of damage from patent trolls. Bessen’s work on automation (see Learning by Doing, Yale 2015), both historical and current, provides a distinct analysis of effects on employment, skills, and wage inequality. Bessen’s work has been widely cited in the press as well as by the US White House and Supreme Court, the European Parliament, and the Federal Trade Commission.

Reviews

[Jim]

I discovered this book from the following blog:
http://nextbigfuture.com/2015/11/valu...
and an hour long video about his ideas,
https://www.goodreads.com/videos/8468...

Reading this as it seems a natural follow-up to the ideas found in Rise of the Robots. Rise of the Robots & this book are both found at the Nashua Public Library. 8)

I am picking this up again after a false start in December. My initial take on this book is the ideas are interesting, relevant and I believe are correct. Although one could easily argue that his conclusions don't necessarily follow his research and observations. There are a lot of factors that lead to rising wages such as unionization, wars,labor shortages in addition to productivity.

A major theme of this book is that the initial introduction of a technology does not automatically equal additional productivity and profit. This comes after a period of learning by doing, changes in infrastructure and increased skills and knowledge of the workforce .This takes time.

For example, weavers on the first power looms could produce 2.5 times the amount of cloth per hour than on hand looms. Over the next 80 years, improvements in the looms and the knowledge and skills of workers generated another 25 fold increase in the output per hour.

Another example , there was an 8 times increase in electricity produced from a ton of coal from 1900 to 1960.

(Not mentioned in book) I believe the advent of fracking and improvements in the process,equipment,oil exploration and worker knowledge have led to improved productivity and cost reductions in natural gas extraction. Thus making more wells viable and creating at least for a period increases in economic wealth in areas of the US.

The following are some notes and fun facts as I read the book

Chapter 1 :

Technology vs Original Invention
* US Textile Mills without a relationship to a strong local machine shop tended to fail.

* Mokyr divided knowledge into the Why and How. How being foremost in the technical innovation implementation.

* James Watts invention as is did not hurl the industrial revolution forward. A series of makers, machinist and working engineers did.
* A common observation during the industrial revolution. The best technology was invented in France and worked out in England.
* Steve Jobs felt it was a "disease" to think a great invention was 90 % of the work

Mass vs Elite Knowledge

* William Gilmore who brought textile machines from GB to the US needed skilled people to show him how to get the machines running.
* at first, Western Mills were 6.5 times more efficient than mills in India and China run by Western foreman. ( no longer this gap)
* technical knowledge is central to the well being of a large number of people.

Knowledge vs Ideas

* Innovation theory considers incentives for invention and largely ignores the incentives for the costs of large knowledge and skill improvements of workers.
* the design of the power loom was easy to replicate but the knowledge to implement it was not.
*incentive in R&D investments and knowledge obtainment are both important in markets.

Dynamic vs Static Technical Knowledge

* early stage of technology- formal technical education harder
*early stage of technology-person to person transfer of info
*early stage of technology-tends to be localized -silicon valley

C2-Skills of the Unskilled

*apprenticeships tended to has small range of skills transfered
*deskilling jobs often create different skill jobs-machine tender, programmers
* textile mill workers invested in education by piece work lower wages.
* his experience with WSIWYG software required much OJT of labor
* learning related costs are sometime 10X cost of softwares/computer systems
* In WW2, US Workers increased ship production 4x faster during the course of the war.

c3-Revolutions in Slow Motion

*a great time to commercialization chart of inventions-pg 39- which shows year patented-years to 1st commercialization-years to shakeout-total years- The mean is 67 years from pattern to maturing..
* Bessemer Process-

[Jonna Higgins-Freese · Rating 3 out of 5]

In a way, this book is an extended version of the argument in _The Breakfast Club_ between The Brain: "without electricity, there'd be no lamps" and The Muscle: "without lamps, there'd be no light."

Innovation is nice, but without the knowledge to implement and use it, no one really gets the benefit of it from a macroeconomic level.

It took me a while to get into this, and some of the Goodreads summaries would have provided about as much information as I'll take from it. One benefit of reading the full book is that it took a while for me to be convinced that the wages of workers are more tied to distribution of knowledge about how to implement technology than unionization and public policy. His argument about the lack of value of unions is flawed: the fact that textile workers wages rose despite the fact that their unions were small might have been due in part to the _threat_ of unionization as well as the skill level of employees, and Bessen doesn't control for that in his analysis (82). I'm still not entirely convinced, but I'm willing to believe that distribution of knowledge is more important than I thought.

The book demonstrates the same sort of respect for manual knowledge Mike Rose's work. Regarding the increase in productivity from experienced weavers: "getting used to the noisy, complicated environment of a weave room with leather drive belts spinning and looms clacking away. Part of it was acquiring specific knowledge such as how to tie a weaver's knot. Part involved practicing manual operations to find faster movements -- for example, to quickly replace an empty shuttle . . . learning how to adjust the tension on the warp so as to minimize broken threads; planning skills, such as coordinating work on multiple looms; and developing the ability to monitor the looms for minor errors and flaws" (25).

Overall, he argues that there has to be some level of standardization of methods and knowledge for workers to develop transferable skills and for wages to rise. For example, he says that graphic designers, whose work has largely replaced that of typesetters, don't earn as much as typesetters did because there are still too many different programs, and they are shifting all the time. (111). Again, though -- typesetters were unionized and graphic designers are not. It doesn't quite make logical sense that workers with widely generalizable skills ought to earn more than those with specialized skills. But perhaps that's the point of the book -- that it's counterintuitive? If so, it's so counterintuitive, and there are so many confounding variables in the data and examples he provides, that I'm not convinced (111).

On the other hand, he did convince me on the subject of licensure -- where safety and health aren't at stake, licensure serves only to keep people from the field -- though his argument here is not that this the depresses wages but that it keeps others from entering the field -- and wouldn't scarce supply increase remuneration?

He did provide the best explanation I've seen for why so much of medicine has been able to be de-skilled: outpatient surgery and minute clinics have standardized certain areas of practice so much that even those without much formal education (LPN's) can gain on the job knowledge that is valuable. Nursing organizations continue to push for more formal certification, however (114)

Some of his arguments just defy common sense: "a higher percentage of working-age Americans are employed than in 1980. The economy has been able to create enough jobs to employ a rising share of [them], despite all the new technology" (118) -- yes, and, more of them _have_ to work precisely _because_ of wage stagnation -- it now takes two to earn what one did before.

He does make a compelling argument against broadening the proportion of those who have college degrees per se; saying some things that it would be nice to hear more loudly and clearly: ""A certified LPN with a year of postsecondary education might be much more employable than someone with a year in a liberal arts bachelor's degree program" 143. At the same time, he doesn't fall into the trap of just being a booster for science and technology degrees: "50% more computer science majors graduate each year than in 1998," yet many grads report jobs are not available. "The demand for college-educated workers is mainly for those who can learn on the job, not for specific vocational skills that they learned in college" (145). At the same time, he argues that grads who are underemployed may simply be those who can't also learn on the job (rather than those who are victims of credential inflation -- see Why Good People Can't Get Jobs: https://www.goodreads.com/review/show...). A deeper look at those confounding variables would have been more convincing to me.

And, as always, my pet peeve: shilling for more vocational degrees by pointing out that vocational jobs often have salaries as high or higher than those of college grads -- without taking into account total compensation: benefits, vacation, etc. This drives me crazy (148). I know a couple with vocational degrees who make what is considered a middle-class household income. But neither has access to decent insurance, so each, separately, has a policy with an annual out of pocket max of $12K and very little preventive coverage. Their income is effectively MUCH less than their salaries, and much less than folks with four year degrees who would be more likely to have access to insurance through work.

I was better able to get on board with his policy analyses:
-- government procurement, by disseminating knowledge among subcontractors and driving toward a common standard, open opportunity and innovation. The examples he gave were in IT in the US and gun procurement just after the Revolutionary War (165). "The gov't backed diverse approaches, architectures, and suppliers." The packet-switched network competed with the telephony circuit-switched network. "Britain . . . [supported] national champions in the computer industry; France backed Minitel, a precursor to the Web produced by its government-owned telephone company" 169. Japan supported its hardware developers such that companies developed their own proprietary software and workers were not able to move from job to job.

--strong patent/IP protections hurt small firms while protecting largeones
--lack of technical knowledge can limit broader adoption of technologies; "in biotech and semiconductors, the personal involvement of the 'star scientists" significantly improves the prospects for successful commercialization. And universities generate much more licensing revenue when faculty members participate and share in the proceeds" (186).
-- employees do better in states that don't enforce noncompetes (189)
-- the ability of young, literate women to work outside the home is predictive of economic success in a variety of contexts (224)
-- tech knowledge to some extent cannot be forced: efforts to run textile mills with slaves failed
-- social institutions that provide support during transitions are important: "Avner Greif and Miurat Iyigun find that England experienced relatively little popular resistance to technology-related economic transformations: they attribute this peacefulness to England's welfare institutions that provided support to the poor. They find a statistical association between the level of welfare spending and innovation. . . . a national ability to implement new technologies depends on how the nation treats the many ordinary people who build and use that technology, and the freedoms, the incentives, and the protections it provides them" (225).

[Myles · Rating 4 out of 5]

This is a pretty useful discussion of the real world impact of work experience and the people who make innovation work. Historically it has taken decades for workers to enjoy the monetary gain produced by technological improvements. But it is nowhere a given that any particular improvement will generate gains in the short term. Worse, the structure of the modern economy is undoubtedly working against the sharing of knowledge, a big way workers get to share in the wealth. Bessen is speaking, of course, of our wages, not those of the average Chinese worker whose wages have taken off in the most recent rounds of technological innovation. Bessen's core argument is that greater access to vocational training in the most modern sense will speed improvements to our wages. At the same time he remarks on some of the trends that will inevitably retard the process. They include the decline of funding for vocational institutions; the restrictions that occupational licensing places on the labour market; the growing tendency of government to favour large corporations over start-ups and universities to innovate; higher standards of secrecy, the growth of non-disclosure and non-compete agreements in hiring and partnering practices; patent laws and the corrosive effect of patent trolls on start-ups, and abusive patent litigation. All of this reduces sharing at the very basic level of industrial organization. Most surprising -- surprising not because it's wrong, just unexpected -- at the end of this study is a stinging indictment of money in politics, certainly American politics. Moneyed interests have stymied patent reform and any encroachment on the territory of licensed practitioners. Bessen's study of the mill workers in the 19th century indicates that there will be a lag in wage growth until standardization prevails in an industry and workers have a chance to catch up to technology. That's a big "until" in this world. If anything, innovation is accelerating these days. Will anybody be able to catch up? Anybody, that is, except artificial intelligence...supercomputers and self-improving bots. If anything, this is a very worrying book. I like to think of this excellent book as an update to William Greider's excellent 1998 introduction to Globalization "One World, Ready or Not: The Manic Logic of Global Capitalism."

[Dalan Mendonca · Rating 5 out of 5]

Outstanding book. 7/5 from me. I think the author does a great job countering common narratives about wages/employment/technology.

Example #1: Yes technology takes away tasks from workers, but that doesn't mean workers get replaced or even get lower wagers. Workers can more by acquiring skills in the restructured industry. Workers don't tend to lose jobs if the overall industry is growing.

Example #2: Education is generally good, but societally we overvalue education; especially in nascent industries where most knowledge is tacit and unstandardized. Somehow a 4-year degree has become a social necessity; but it provides general skills that don't lead to higher wages. Instead the focus should be on vocational education where workers acquire skills directly connected with job functions.

Examples #3: Patents are seen as protecting IP but repeated pattern is that when a technology is nascent people share knowledge to improve the technology. Only when a technology matures does knowledge sharing decline. So technology policy has to evaluate whether technology is mature or nascent.

Overall, the book presents a complex picture of the interactions between technology, workers, wages, wealth and national prosperity. My only desire is that the author go one step further and formulate a Porters 5 forces style model of how these factors interact so that the overall picture can be understood better and we can create better policy based on it.

This book is right up there with Carlota Perez's Technological Revolutions and Financial Capital in explaining broad phenomena in a technology rich society.

[Johnny · Rating 5 out of 5]

After hearing James Bessen interviewed on the EconTalk podcast, I had to order an eBook version of Learning by Doing: The Real Connection Between Innovation, Wages, and Wealth. First of all, I found Besson’s argument that public policy has been heading the wrong direction as opposed to policies which would encourage acquisition of new skills. Some have called the U.S. a “meritocracy” and there was a certain amount of truth to that in the past. Bessen cites economists as believing that “China will not be able to sustain its economic growth unless it provides greater political democracy.” (p. 226) By this, Bessen means “opportunity” as part of the economic structure. It was surprising to me to read, “Counties in England that provided greater welfare payments had fewer food riots but also more patents; countries in Europe that provided more welfare also displayed more technological innovations at international exhibitions.” (p. 225)

Of course, measuring progress by patents may not be the best metric. “When new technology coexists with inferior alternatives, patents have limited value.” (p. 182) Bessen argues that patents inhibit emerging technologies, but after standardization and the maturation of technologies, they serve the purpose of protecting the patent holders who have invested heavily in them. Yet, today, patent law is a mess and “patent trolls” buy up obscure and ambiguous claims in order to kill start-up companies which cannot afford litigation to protect their own research. Indeed, it is estimated, “…patent trolls impose a 10 to 20 percent tax on R&D;” (p. 196). I love the quotation from Congressman Jim Cooper, “The past, in general, is over-represented in Washington. The future has no lobbyists.” (p. 214) In a 2008 survey, 61% of tech start-up firms had never applied for a patent; in software, 76% had never applied (p. 197).

In a similar vein, Bessen argues that occupational licensure (now over 29% of all jobs) serves primarily to protect the priesthood (those already “ordained” into the occupation) rather than to protect the public. “Occupational licensing restrictions increase wages by 18 percent on average.” (p. 157) That’s good for the “priesthood,” but restricts job growth and vocational training (p. 213).

This is the second book I’ve read where the author points out that job mobility in California as opposed to “non-compete” agreements in Boston created an environment of knowledge sharing and a more productive, vibrant industry (pp. 188-189). Most importantly, Bessen draws from the development of the textile industry (and even steel industry) to illustrate how job mobility contributed to knowledge sharing in those industries. For me, having covered the so-called “convergence era” in the 1990s, one poignant note stated “Time Warner did not have a strong incentive to establish open standards in the development of interactive TV.” (p. 167)

Further, Learning by Doing: The Real Connection Between Innovation, Wages, and Wealth helped me understand the U.S. Labor Movement of the late 19th century. Bessen explained, “U.S. mill owners argued that their workers were superior to the British because of their education. An examination of the payroll records from the mills shows that the illiterate weavers were indeed less productive. Literate new hires learned the required skills faster and better (p. 139). Later, greater turnover made it harder for mill owners to recoup their investments in literate weavers (p. 140).

Learning by Doing: The Real Connection Between Innovation, Wages, and Wealth has fascinating policy ideas (funding community and vocational colleges better than the current bias toward four-year colleges, reforming patent law, and avoiding non-compete agreements), but also cites information designed to keep people from being afraid of automation. For example, “While some banks saw ATMs as replacements for tellers, others found greater success using tellers as part of the “customer service team” that sought to draw customers to more profitable services, such as investment management.” (p. 108) The technical term is relationship banking (p. 109).

Learning by Doing: The Real Connection Between Innovation, Wages, and Wealth has already given me lots to think about and I’m sure I will be using it for reference and continued discussion in the future.

[Alok · Rating 4 out of 5]

This book covers the surprisingly indirect connection between productivity and wage growth. The historical evidence centers around the transition from hand weaving to power looms. There was a gap of nearly 70 years between the introduction of the power loom and the increase in wages, during which time the owners of the mills reaped outsize profits. The key factors allowing increase in wages were that the work became standardized enough that workers could easily move between mills, thus forcing the factory owners to pay more. Also, there were competing forms of non-mill work which allowed workers to exit the mill labor force. The important thing is the we could be in for a few generations of slow wage growth (perhaps we're already one generation in).