The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, by Erik Brynjolfsson and Andrew McAfee

New York: W. W. Norton & Company, 2014, 0 393 23935 5, hbk, 320 pp, $26.95. Audio edition: Grand Haven, Michigan: Brilliance Audio, 2014.

This book was recommended to me as a technology-based argument for the basic income guarantee (BIG), and it is, but its support is tentative and only for BIG in the form of the Negative Income Tax (NIT), not in the form of a Universal Basic Income (UBI).

The authors define the computer revolution that is currently underway as ‘the second machine age’. The industrial revolution was ‘the first machine age’. It brought machines that could apply power to do simple but profoundly important tasks, eventually replacing most human- and animal-powered industries with steam, electrical power, and so on. Machines of the first machine age could often do those tasks much better than humans or beasts of burden ever could. For example, the replacements for horses – automobiles, trains, and airplanes – can carry more people and more cargo further and faster than horses ever could.

Machines of the second machine age have gone beyond the application of power; they are also replacing some human brainwork. Calculators have been around so long that few people are aware they replaced a form of human labour, called ‘computers.’ In the early 20th century, ‘computers’ were people who did computations. It was skilled brainwork, far beyond the capabilities of the up-and-coming technologies of the day, such as the internal combustion engine. Computers (as we define the term today) have almost entirely replaced that form of human labour, and their ability to substitute for human labour only continues to increase – especially when combined with robotics.

The computational powers of computers are so strong that they can already beat the best chess masters and ‘Jeopardy’ champions. Self-driving cars, which have turned driving into a complex computational task, will not only relieve us all of the task of driving to work, they have the potential to put every professional driver out of business. Perhaps computers, then, will someday learn not just to calculate, but also to think and evaluate. If so, might they eventually replace the need for all human labour?

Perhaps, but Erik Brynjolfsson and Andrew McAfee, the authors of The Second Machine Age, do not base their arguments on any such scenario. The possibility of a truly thinking computer is out there, but no one knows how to make a computer think, and no one knows when or how that might happen.

So, the authors focus on the improvements in computers that we can see and envision right now: machines that can augment and aid human thought with computational ability increasing at the current exponential rate. As long as computers are calculating but not truly thinking, humans will have an important role in production. For example, although computers can beat an unaided chess master, they cannot beat a reasonably skilled human chess player aided by computer. This is the focus of the book: computers and robotics taking over routinized tasks (both physical and mental), while humans still perform the deep thinking with access to aid from more and more computer power.

This change will be enough to transform the labour market radically and eliminate many (if not most) of the jobs that currently exist. At the enormous rate of increase in computing power, one does not have to envision a self-aware, sentient machine to see that the effects on the economy will be profound. According to the authors, ‘in the next 24 months, the planet will add more computer power than it did in all previous history; over the next 24 years, the increase will likely be over a thousand-fold’.

The authors’ analysis of those changes is very much based on mainstream economic theory. In the book’s analysis, increases in unemployment and decreases in wages are attributed almost entirely to a decline in demand for labour thanks to the introduction of labour-replacing technology. Political economy considerations, in which powerful people and corporations manipulate the rules of the economy to keep wages low and employment precarious, are not addressed. When the authors consider shifting taxes from payroll to pollution, they don’t consider that powerful corporations have been using their power over the political process very effectively to block any such changes.

Yet, the book demonstrates that even with purely mainstream economic tools, the need to do something is obvious. We have to address the effects of the computer revolution on the labor market. The second machine age creates an enormous opportunity for everyone to become free from drudgery, to focus their time on the goals that they care most about. But it also creates a great danger in which all the benefits of the second machine age will go to the people and corporations who own the machines, while the vast majority of people around the world who depend on the labour market to make their living will find themselves fighting for fewer jobs with lower and lower wages.

The technology-replacement argument for BIG has been a major strand in BIG literature at least since Robert Theobald began writing about the ‘triple revolution’ in the early 1960s (Mostly in three works, The Challenge of Abundance (1961), The Triple Revolution (1964), andThe Guaranteed Income (1966)). So, approaching this book as I did, I was on the lookout through a large chunk of the book, waiting for BIG to come up. I was very surprised to see the entire ‘Policy Recommendations’ chapter go by without a mention of BIG.

The authors finally addressed BIG in the penultimate chapter entitled, ‘long-term recommendations’. In the audio version of the book, the authors spend about 20 minutes (out of the 9-hour audiobook) talking about BIG. They recount some of the history of the guaranteed income movement in the United States with sympathy, and write, ‘Will we need to revive the idea of a basic income in the decades to come? Maybe, but it’s not our first choice.’ They opt instead for an NIT, writing ‘We support turning the Earned Income Tax Credit into a full-fledged Negative Income Tax by making it larger and making it universal.’

Their discussion of why they prefer the NIT to UBI is perhaps the weakest part of the book. They favour work. They want to maintain the wage-labour economy, because, taking inspiration from Voltaire, they argue that work saves people from three great evils: boredom, vice, and need. I am sceptical about this claim. I view it as an employers’ slogan to justify a subservient workforce, but my scepticism about this argument is not why I find the book’s argument for the NIT over UBI to be the weakest part of the book. The reason is that the argument from work-incentives gives no reason to prefer the NIT to UBI. The authors view the NIT as a ‘work subsidy,’ but it is no more a work subsidy than UBI.

The NIT and the UBI are both BIGs, by that, I mean they both guarantee a certain level below which no one’s income will fall – call this the ‘grant level’. Both allow people to live without working. UBI does this by giving the grant to everyone whether they work or not, but taxing them on their private income. NIT does this by giving the full grant only to those who make no private income and taking a little of it back as they make private income. In standard economic theory, the ‘take-back rate’ of the NIT is equivalent to the ‘tax-rate’ of the UBI, and so either one can be called ‘marginal tax rate’.

Applying standard mainstream economic theory (which is used throughout the book), the variables that affect people’s labour market behaviour are the grant level and marginal tax rate. The higher the grant level and the higher the marginal tax rate, the lower the incentive to work whether the BIG is an NIT or a UBI. You can have an NIT or a UBI with high or low marginal tax rates and grant levels, and you can have a UBI or an NIT that have the same grant level and marginal tax rate. It is for this reason that Milton Friedman, the economist and champion of the NIT, gave for drawing equivalence between the two programs:

INTERVIEWER: ‘How do you evaluate the proposition of a basic or citizen´s income compared to the alternative of a negative income tax?’
FRIEDMAN: ‘A basic or Citizen’s Income is not an alternative to a negative income tax. It is simply another way to introduce a negative income tax.’
(Eduardo Suplicy, USBIG NewsFlash interview, June 2000,

If the book’s arguments for work incentives are sound, I see an argument for a modest BIG with a low marginal tax rate, but I see no argument one way or another why the BIG should be under the NIT or the UBI model.

Whatever one thinks about the issue of NIT versus UBI, the book presents an extremely sophisticated and powerful argument for moving in the direction of BIG. Therefore, it is a book that anyone interested in any form of BIG should examine closely.