Tyler Cowen has followed up the best-selling The Great Stagnation with another profoundly interesting book on the impact of new technology on the economy and beyond. Average is Over: Powering America Beyond the Age of the Great Stagnation picks up the theme of the disruption technological change and the consequent restructuring of business and focuses on the effects on individual jobs and what people might be able to do to safeguard their livelihood.
The bottom line is that you need to think about which of your skills can easily be done – better, cheaper – by a computer, and work hard at developing other skills. The general message is relatively optimistic, with the book concluding that the combination of machine plus human power being much more productive than either alone; one of the examples of how and why this is so is ‘freestyle chess’. So a computer able to do fast routine work plus a human who has either advanced cognitive skills or non-cognitive skills that computers can’t attain could prove a winning team. The spread of computers through the economy would in that case prove to be another instance, seen many times in economic history, of investment in capital ultimately increasing the productivity and living standards of the people working with that capital.
However, the more specific message of the book is rather sober: Cowen doubts that all that many people in America at least have the focus and capacity for work and concentration to turn themselves into complements for robots rather than being substituted by them. And so far there is certainly plenty of evidence that routine work in many sectors of the economy, in the middle-skilled, middle-income bracket, is being away by computerisation, either directly through automation or indirectly through offshoring. About 60% of the jobs lost during the US recession have been in mid-wage occupations. Wages for the median male worker declined by about 28% between 1969 and 2009, this with no nuclear war, no asteroid striking earth, or other disaster. Nor is this disappearing middle of the income distribution just a US phenomenon: last week’s report on UK incomes from the Resolution Foundation pointed to similar evidence. Cowen writes: “The obvious and direct beneficiaries [of ever-more powerful computers] will be the humans who are adept at working with computers. … That means humans with strong math and analytic skills, humans who are comfortable working with computers because they understand their operation.” He argues that the scope of the phenomenon in the wider economy will only grow, pointing to driverless cars and taxi driver jobs, for example. Or think of those dreadful machines that are replacing supermarket cashiers.
A section of the book uses chess as an illustration of the trends – what computers can do, and can’t, and what human skills are substitutes and complements for computer skills. In short, computers do not have intuition, and cannot actually move chess pieces on a board. Nor are people interested in watching computers play chess, nor even human-computer teams – they want to watch other people play, and thanks to the internet chess has become a mass spectator, global sport.
There is a very interesting section on ethics – I wished it had been longer. Driverless cars will be more efficient. When the investment in the infrastructure is made, they will economise on drivers, be able to use fuel more efficiently, observe speed limits, probably be much, much safer. But how will they respond to moral dilemmas? If the choice is running over a baby carriage on a crossing by swerving into a group of five pensioners, what will have been coded into their programmes? Another section of the book predicts that the detailed tracking of measurements will transform all kinds of services, such as medical care. Patients will have increasingly detailed metrics on individual doctors’ performances; doctors will have detailed measures of how patients behave – do they take the full course of medicine? Do they make time-wasting visits? Do they smoke? It will be hard for people to make the judgements weighing up the things that can be measured easily (like death or recovery rates) and those that can’t, such as a caring personality, or a willingness to treat the most serious diseases. Steadily, without discussion, a moral framework is becoming encoded into the machines around us, and by the habits of use that are developing.
I’m not sure I share the book’s view about exactly what kinds of jobs will be left to humans. Cowen argues that these will be marketing jobs, work synthesising lots of areas of expertise to help people make complicated judgements, and coaching jobs for those lucky individuals who can develop their complementary talents, whether that’s being a top doctor, a top tennis player or a top executive. A more optimistic perspective would see opportunities for the growth of fulfilling creative work, such as acting or gardening or creating beautiful clothes; let the robots do the dull, repetitive jobs. Whether that is possible depends of course on whether the productivity gains from computerisation can be shared, and whether income can be distributed from the owners of the machines to the wider society. Clearly, that has not happened during the past generation.
All capitalist societies have faced this issue, however. How can people who are less productive (because there is less capital linked to their job, whether that’s machine-embedded capital or human capital) not be left so far behind that it becomes socially unsustainable. In past centuries that involved, for example, the spread of public schooling to primary and then secondary and tertiary levels, the growth of unions campaigning effectively for labour’s share of profits, redistribution through progressive taxation paying for public services available to all, and so on. I’d have liked more in Average is Over about how social and political institutions might respond to the challenge of the robots, more political economy. Maybe that will be his next book. Meanwhile, there is a clear message here for every reader: if you want to win the race against the machines, you’d better start running.