Monday, 9 April 2012

Does technology cause inequality?

I'm not thinking of the psychic damage caused when you don't have the latest Apple gizmo, but the way in which technological change is managed and the consequences this has for income equality.

While I was musing recently on the proposal to revamp ICT education in schools, I came across a number of papers discussing the relationship between education and wages over the last 60 years. These broadly show that income equality increased as the post-WW2 reform programmes extended education to the working classes, stagnated as these programmes ran out of steam in the 70s and 80s, and went into reverse thereafter. The causal explanation for this correlation is technology:

"Technological change increases the relative demand for skilled and educated workers, but educational advance increases their relative supply. This “race” between education and technology can produce rising, declining, or stable levels of economic inequality" [para 3]

The economist Daron Acemoglu has observed that whereas technological change in the 19th century was largely skill-replacing, i.e. substituting unskilled for skilled workers, technological change in the 20th century has been skill-biased, i.e. requiring higher skilled workers and thus placing a premium on education. One reason for this change may be the impact that the available labour supply has on the opportunity for profit growth. Where you have a lot of cheap, unskilled labour, skill-replacing technologies are attractive as they allow you to push down on labour costs. Where you have a lot of skilled labour, technologies that exploit those skills are attractive as they can boost profit through increased productivity.

"In contrast, the twentieth century has been characterized by skill-biased technical change because the rapid increase in the supply of skilled workers has induced the development of skill-complementary technologies." [pg 64]

New technologies spur the demand for education, while increased education spurs the development of new technologies. This sounds like a virtuous circle, so why should this give rise to increased inequality? The conventional answer would be that education has raced ahead of technological change, so we have an over-supply of educated workers, however this doesn't square with the obvious acceleration of technological change over the last 20 years, particularly in terms of computing and communications. Indeed, we know there is a dearth of sufficiently skilled workers because of the high wage premium for techies since the 80s.

In fact, technological change over the last 30 years has seen a mixture of different dynamics at work. Skill-replacing technologies did not end with the 19th century and have plenty of mileage yet, while offshoring allows skill-biased productivity gains to be combined with skill-replacing cost reductions. It is onshore office jobs where the virtuous circle is found, though even here it is limited to a core, with jobs at the margin vulnerable to automation and offshoring. At this point it is worth distinguishing between technical skills (often learnt on the job) and general knowledge (learnt at school and college):

"In periods of technological change, those who are relatively invested in technology-specific skills find it harder and more costly to adapt to the new changes. Conversely, workers who are heavily invested in general skills, such as education, adapt to the new technology more easily and less costly ." [pg 31]

It's easy to think of examples where a new technology has disrupted a sector of the skilled labour market, but it's not so obvious why education generally would make you more adaptable. The reason may be, to extend Acemoglu's thinking, that recent technological change has been biased towards the aptitudes of the educated.

At one level this is easy to dismiss because major technological breakthroughs are not usually driven by the needs of office workers, however when you distinguish between technology platforms (or macroinventions) and specific applications (or microinventions) then it looks more plausible. Thus the academic research into packet-switching that led to the Internet soon produced applications that replicated the form of memos (SMTP), bulletin boards (Usenet) and citations (HTTP), which office-based academics were familiar with.

What has been notable is the degree to which these new technologies have given rise to activities that have soaked up the productivity gains. Thus emails are more effective than paper memos, but this has led to everyone writing more with no overall reduction in labour time. The added speed of research that the Web has brought has been offset by time spent on "capricious browsing". The clamour for social media's use within business has long looked like a solution in search of a problem.

Schools and colleges are meant to "prepare" kids for the "world of work", i.e. they have an instrumental function (or at least a signalling value) beyond the intrinsic value of learning. Elementary and secondary modern schools used to focus on the discipline and obedience deemed necessary for factory fodder. Private schools cultivate self-confidence and an assumption of superiority. Modern secondary schools reflect modern business practices in the use of projects, teamworking, self-study and group tutorials (i.e. meetings). The actual skills we're teaching them may be "how to behave in an office", in which case learning how to use MS-Office is perhaps not as witless as it seems.

This may explain the paradoxical fall in computer science student numbers at a time when the wage premium should encourage uptake. The lesson of the last twenty years is that success in white collar jobs is largely about being a generalist, not a specialist, getting into management (of people or things) and generally making yourself look busy. When the most popular employers among graduates include business and accounting consultancies, then you know you're dealing with a generation bereft of romantic illusions.

Technology does not directly cause inequality. Instead, inequality is the consequence of the unequal way in which technology is applied to jobs. It is essentially skill-replacing at the bottom end of the income scale, enabling automation and offshoring of both skilled and unskilled roles. This is a return to 19th century practice. Indeed, the mid-20th century now looks like an anomaly: the demands of war and reconstruction, together with the limitations of transport and communication, pushed up the value of domestic workers, both skilled and unskilled, leading to reduced inequality.

At the top end of the scale (i.e. median earnings and above), the application of technology is skill-biased. Specifically, it is biased towards the skills and behaviours acquired by the educated, those increasingly referred to as "knowledge workers" (as if manual workers are barely sentient). Additionally, the self-directed mode of these skills and behaviours allows the productivity gains achieved by the technology to be dissipated, thus preserving white collar jobs. This waste of productivity at the top end is tolerated by business leaders so long as the beneficiaries aggressively pursue productivity gains through skill-replacement at the bottom end.

I suspect the next major technology wave, or macroinvention, will coincide with the near exhaustion of bottom-end skill-replacement and will lead to net productivity gains among white collar staff. The squeezed middle 'ain't seen nothin' yet.

No comments:

Post a Comment