It seems self-evident that information technology makes us more productive.

However, academics and technology sceptics know all about the productivity paradox. It was popularised in a 1993 article by Professor Erik Brynjolfsson, who noted an apparent contradiction between advances in computing power and the slow growth of productivity.

Let’s face it; technology’s transformative effect on business isn’t always apparent. Be brutally honest, even with all of your new devices and mobile working capabilities, how much more productive are you today than you were 10-15 years ago? Enough to justify every dollar you’ve spent on technology in the interim?

More likely your productivity is being spread more thinly over what was previously your leisure or commuting time. This is corroborated, in part at least, by responses to a Clarian Human Resources study completed in conjunction with Massey University this year, the Great New Zealand Employment Survey 2013 (a nationwide online survey based on 334 responses). For example, one comment read: “Constantly accessible. People expect faster responses. Too easy to keep checking in on emails at home and on leave”. Respondents cited “excess workload” as a barrier to performance and nearly two-thirds of respondents felt IT led to spending more time on work.

It’s partly a perceptual problem, of course; just because you’re at work doesn’t mean you’re working. As the Economist’s contributor Buttonwood wryly noted, the ability to watch funny cat videos doesn’t count as increased productivity, and the same publication has been saying new technologies don’t automatically lift productivity since at least 2003: “Firms need to work out how to reorganise their business to make best use of any important new technologies before they can reap the full rewards.”

Back in 2000 Professor Robert Gordon at Northwestern University in the US wrote a paper (Interpreting the ‘One Big Wave’ in US Long Term Productivity Growth) in which he asked whether the computer and internet revolutions are as important as the first industrial (steam) and second industrial (electrical and internal combustion) revolutions. He contended many of the inventions that initially led to the deployment of computers occurred in the 1970s and 1980s and since then the majority of notable developments had been in communications and entertainment.

“But that was before the effects of the internet,” you reasonably respond. “Now we have the cloud, cheap storage, reliable web search and robust, free email!” So why aren’t these innovations unambiguously reflected in our productivity figures? Well, for one thing, says Oxford University economist Paul David, by comparison there was no notable productivity growth until at least 40 years after the introduction of electric power. It took until around 1920 for US machinery to be connected and for organisations to re-engineer themselves for the benefits of electricity. David also calculates that a technology only begins to significantly affect productivity when it has reached a 50 percent penetration rate. US computer use, for example, only reached this mark in around 2000.

More input, less output
The Australian government’s Productivity Commission is studying the problem of manufacturing’s contribution to the decline in productivity growth and undertaking work to identify its causes. Its findings aren’t expected until after this edition goes to print. But in its June 2013 National CEO Survey, the Australian Industry Group says the fact Australian businesses have been keen technology adopters over the past two decades has had a positive impact on productivity at a company level. To maintain this momentum, it concludes, significant new policy initiatives are required, including the development of a national workforce skills strategy for the digital economy.

Meanwhile, a September 2012 paper by the New Zealand Productivity Commission, Productivity by the numbers: The New Zealand experience, finds New Zealanders working more hours but producing less than workers in other countries: “New Zealanders work about 15 percent longer than the OECD average to produce about 20 percent less output per person,” the paper says. New Zealand’s labour productivity has been falling behind other OECD countries for decades, it seems. Productivity commissioner Murray Sherwin has also said that even though the country has invested heavily in ICT technology, many organisations have failed to turn their investment into meaningful productivity growth.

Healthy scepticism
Some digital commentators are sceptical about the metrics traditionally used to measure productivity. Geof Heydon is director of business development for the information sciences group at the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in Sydney. “If you take something that’s been happening for 100 years and try and work out how much it’s improved, there’s a fairly good chance that you won’t find massive improvement,” Heydon cautions.

However, he is dubious about research findings suggesting technology isn’t increasing productivity. “More often than not those research programmes are run by various big consulting companies, and quite often the outcomes are very much dependent on who’s funding the research. The research required to get proper answers on these things is extensive and mathematically complex.”

Underscoring the difficulty of measuring productivity in the digital age, Heydon says it’s challenging to analyse an entire economy and understand what’s happening to it from a computing perspective. “It’s easy to miss critical aspects of what’s digital and hard to be completely thorough and detailed about it.”

Technology timeline
Pinpointing the source of productivity gains when new forms of power were introduced was relatively straightforward but seems to have become more elusive with each progressive innovation. For example, from the mid-1880s, electrification dramatically increased the productivity of factories. Also, the centralisation of electricity generation meant more businesses could afford electricity because they paid only for the power they used.

“If you take something that’s been happening for 100 years and try and work out how much it’s improved, there’s a fairly good chance that you won’t find massive improvement.”
Geof Heydon, director of business development, information sciences group, CSIRO

The earliest information technology hardware took the form of adding machines and unit recording equipment, which processed data by running punched cards through tabulating machines. Comparable human calculations required more manpower and were subject to greater levels of human error. The first completely transistorised calculator dates back to 1955 when IBM introduced its 608 machine (those punched cards weren’t fully superseded until the 1980s).

The lack of portability of early electric typewriters in the early 1900s is one reason why they took time to realise sizable productivity gains and they didn’t wholly supplant manual machines before the very first word processors began to sweep both typewriter variants aside in the late 1970s and early 1980s.

…