The Second Machine Age (Norton 2014) or 2MA by Erik Brynjolfsson and Andrew McAfee has received several excellent reviews. Some can be located via Amazon or other web searches where summaries of the book are also available. Rather than repeat them, I note aspects of 2MA which especially struck me.
- As a book about economic history and the economic aspects of current technological developments, it is written in a style which is widely accessible. There is an absence of economic jargon and the explanation of economic concepts is understandable to the informed layperson. Even the difference between mean, median and mode is outlined, for example when dealing with the measurement of changing income inequality.
- The book’s title identifies around 1765 as the start of the first machine age with the development of Watt’s steam engine, which substituted mechanical power for man and animal (horse) power for many types of economic activity. Railways were a big part of this age. Water power and canals were also features of this earlier period. The latter are now often used for tourism, although some like the Suez and Panama remain as busy shipping highways.
- The second age, 2MA, relates to computers and developments in digital communications. Like most inventions they begin with a series of inputs, one of which was Alan Turing who is often considered the founder of the modern computer with the publication of his 1936 paper. But things took off later. A 1965 article by Gordon Moore, then working for Fairchild Semiconductor, predicted correctly that “Integrated circuits will lead to such wonders as home computers – or at least terminals connected to a central computer – automatic controls for automobiles, and personal portable communications equipment.” Home computing and laptops appeared in the early 1980s. This paper was the source of Moore’s Law, which forecast that the amount of integrated circuit computing power bought for one dollar would double each year. That has happened for over four decades and some extend this forecast for another eighteen months. Others say the law will end in about 15 years due to various physical constraints. Who knows? But scientists are probably better forecasters than economists and meteorologists.
- Government statistical agencies first noted information technology as a corporate investment expenditure in 1958, another approximation for when 2MA begins. The world is now in its early stages with pioneers like Bill Gates (Microsoft), Steve Jobs and Steve Wozniak (Apple), Jeff Bezos (Amazon), Mark Zuckerberg (Facebook), and Larry Page (Google). Not long ago the leaders were persons in charge of firms like IBM, Hewlett Packard, Digital Equipment, Dell and Cisco. Because this 2MA is still young, the exponential changes are best graphed logarithmically showing the growth of supercomputer speeds, supercomputer energy efficiency, residential internet download speed and hard drive cost efficiency, as well as the number of microprocessors per chip. A graph with a normal scale would go off the top of the page.
- Two schools of thought prevail about the prospects for future growth. One, supported by Tyler Cowen, sees economic growth declining in the future because we have “picked the low-hanging fruit” of recent technological changes. Another, supported by the authors of 2MA, argues that we are in the early stages of a new machine age and that there will be many applications of information technology which entrepreneurs will introduce. They also argue that the way we measure economic output in terms of GDP grossly underestimates the actual output of the economy, and that new diagnostic tools are required. They do agree that the occupational structure of the labour force has changed and that this accounts in part for growing income inequality.
- What are the shortcomings of the system of national accounts which provides a measure of annual and quarterly GDP? The US accounts were developed in 1937 by a NBER team lead by Simon Kuznets, and have since been refined and adopted by other countries. Economic texts have always noted problems (they are listed in my 1985 fifth edition of Lipseys’ introductory text and I am sure it would have been in the first edition). National accounting does not distinguish between activity associated with cleaning up after an oil spill and with producing tankers. Both are counted as part of annual GDP. And there is nothing like a war to expand economic output. As well, some economic activities go unreported such as criminal acts and certain activities including the services of housewives. As a measurement of change, as opposed to levels of economic activity, if these omissions are constant, then it doesn’t matter too much. But with 2MA the changes are significant.
- Much of what is consumed each year is now either free or cheaper than in the past. Take Wikipedia, it is available for free over the internet, except for the cost of having an internet connection. Individuals provide and consumers use the content for no charge, leaving no recorded activity for the national accounts. Many other electronically delivered services have one or other or both of these features. One result is that calculations of ratios such as the productivity of labour (output per worker) would be higher if the numerator (output) was fully reported in monetary terms. A similar argument could be made about many of the apps which are prepared, often for free, to access some service which may or may not valued in monetary terms. (Chapter 8 of 2MA elaborates on these issues).
- Contrast GDP accounting and reporting with other activities which do record flows of activity. Stock market prices are available in real time, as are current air, rail, and probably truck transportation data (unless the plane disappears). The same is true for the weather, and medical conditions of a patient who wears some type of sensor for blood, heart and other conditions. Why is this not possible for the output of the economy which is used for setting economic policies? Where is the app for real time GDP measurement? (In Canada, Statistics Canada releases information on a daily basis, but not for the day but for some past period. For GDP, the latest data are for January 2014 as accessed in The Daily for April 4th, 2014.) One explanation is that it takes time to collect the data. But if the data are being generated continuously, then it should be possible to collect and report them even if they are revised at a later date – which is the present case anyway.
A future posting will examine what 2MA means for the nature of employment, education and training, and income inequality.