Riffing With Rishi

The pandemic has caused a re-evaluation of a number of assumptions about how society should be organised. I'm going to focus here on the voguish topic of working from home (WFH). Outside of the financial sector and public spending, the economic intervention of the state has largely been rhetorical since 2008 and you could be forgiven for thinking that the more recent enthusiasm for "levelling-up" is merely intended to compensate for the class bias of quantitative easing and austerity. But while it seems unlikely that the North of England will blossom as a result of some free-ports and the relocation of a few Whitehall functions, there is reason to believe that the government will play a crucial role in determining whether WFH becomes the new normal for many people. Rishi Sunak's suggestion this week that many workers want the interaction of the office environment suggests that they are not keen to facilitate too much change. But what interests me is not his implicit support for commercial landlords but the explicit idea that offices aid productivity through some combination of "riffing", "spontaneity" and "team-building". 

Low or stagnant productivity growth has been an issue in developed economies since the 1970s, which coincides with a period of rapid growth in whitecollar roles and new office buildings. In 2019 there were more people in the UK working in offices, both as a quantum and as a percentage of the workforce, than at any time in history. Some attribute this record of poor productivity growth over 50 years (albeit with a short interlude in the 1990s) to the failure of IT to replicate the productivity boosts of earlier general purpose technologies (GPTs), like steam power or electric motors. Thus the Solow paradox from 1987: "You can see the computer age everywhere but in the productivity statistics". The question of whether IT wasn't as impactful as previous GPTs, or whether the benefits were simply lagging, appeared to have been answered when productivity growth increased during the last decade of the century, suggesting that businesses had finally worked out how to use those computers. 

Others see declining productivity growth as an artefact of the neoliberal reorganisation of production, and not necessarily something to worry about. The theory is that globalisation simply shifted productivity growth offshore. The developed economies then moved more activity from production to consumption. Support for this theory was provided by the relatively high rates of productivity growth in domestic manufacturing even as it shrank as a proportion of the economy. But it is easy to mismeasure productivity in services, particularly because improvements in quality (the utility to the consumer) cannot be captured as straightforwardly as improvements in quantity. Manufacturing can be more easily compared across firms because they use common technologies. This in turn means that there is less variation in productivity growth: when a new technology appears it is quickly disseminated. In contrast, services exhibit a much wider range of productivity, reflecting greater variations in skill, market structure and intellectual property.

The return to low productivity growth in the 2000s has been explained through the same two mechanisms. In terms of technology, the low-hanging fruit of the IT revolution, such as improvements in supply-chain and backoffice functions, were one-off gains that were quickly exhausted once fast Internet access was deployed in the 1990s, while the distractions of social media since have shown how counter-productive technology can be. IT investment since the millennium has tended to focus on retail rather than production (apps, online buying), which means that productivity gains are more marginal and those that are realised may disappear at the aggregate level because they are distributive (e.g. gaining market share at the expense of a competitor may not translate into increased sectoral output or lower firm production costs). In terms of the organisation of the economy and trade, the reshoring of high-productivity manufacturing has been too slight to redress the impact of globalisation, while growing wealth inequality has continued to feed low-productivity services (from home deliveries for the middle-classes to concierge services for the rich).

It's true that previous GPTs took many years to fully penetrate industry and deliver their benefits, though these lag-times do appear to have speeded-up. For example, the internal combustion engine became prevalent much more quickly than steam engines. It's also the case that legacy GPTs tend to stick around. Consider how steam, diesel and electric trains were all in simultaneous use in the first half of the twentieth century. The reason for the cycle of slow adoption and reluctant abandonment is that work and infrastructure needs to be fundamentally reorganised around the new technology (the 1990s was the era of process engineering, which is what drove many productivity gains). As the optimum arrangement is only discovered through trial and error, and as the changes require large investments, this is inevitably a slow process. In comparative terms, the two decade lag in the realisation of IT's benefits was quite short. This was probably due to its easier substitutability in certain sub-processes (e.g. email replacing paper post) and its rapidly falling unit costs combined with rapidly increasing performance (Moore's Law etc).

Part of the attraction of the shift to WFH in the wake of the pandemic for economists and social scientists is the expectation that it could be both significant in scale and relatively sudden. In other words, we might be able to see the impact on productivity within a few years. Another factor is the belief that a lot of the work on optimisation has already been done by those who pioneered remote working following the deployment of broadband in the 1990s. Video-conferencing is now a mature technology while many firms had already migrated their core systems to the cloud before Covid-19 appeared. The optimistic school of thought holds that all the conditions are in place for a paradigm-shift in working practices. Noah Smith is representative of this but also of the assumption that these changes can occur organically as a result of market forces. Amusingly, some of his vision revives past predictions that never materialised: that we can all become digital crofters; that firms can outsource individual tasks (neither Amazon's Mechanical Turk nor TaskRabbit have proved revolutionary); and that the boundaries between firms will become increasingly blurred (the Coasian ideal).

One area Smith is silent on is labour organisation. He correctly notes that the productivity gains of electricity in manufacturing came with factory reorganisation, from belt-driven machines powered by large steam engines to workstations driven by local electric motors powered through wiring. This "allowed workers to do things when they needed to be done instead of adjusting their workflow to the rhythm of a giant machine". But it also allowed workers to "go slow" or call "lightning" strikes. The twentieth century pattern of industrial relations was set by variable-speed motors and the ability to cut electricity at the flick of a switch. There is always a labour dimension. In an earlier GPT cycle, the move from putting-out in the textile industry to water-powered mills not only concentrated production it concentrated workers and transformed them into wage labour: the traumatic process of proletarianisation. WFH may have few labour organisation implications for managerial and professional staff, but what if it is extended to unionised sections of the labour force? 

WFH potentially allows firms to forgo or reduce some costs, such as office space, but it also allows them to extend exploitation of the worker, notably in the use of the home as a workplace. This is unlikely to be compensated for, not least on the grounds that the worker gains by not having to pay for commuting. In other words, capitalism's historic ability to transfer the burden of various costs onto labour hasn't isn't going to change. In simple financial terms it may be about to get gradually worse because housing costs have historically risen faster than transport costs. In a putting-out or piece-rate system the worker usually provides both workplace and tools. As WFH has become more common and more homes have personal computers and Internet connections (if not national broadband), the expectation has grown that the worker will provide and the employer will at best subsidise. This suggests a trajectory in which WFH and the gig-economy increasingly converge (a model being sold as "free-agent entrepreneurship" in some quarters).

One reason why employers are increasingly sanguine about remote working is that they can see the potential for greater work surveillance through technology. This isn't about preventing skiving, as workers with minimal control over their jobs will ultimately be paid on the basis of their measured productivity anyway - e.g. the number of transactions completed or cases processed. This is more about disaggregating the workforce so that productivity can be optimised at the level of the individual. As such, it continues the trend of personal performance management that emerged in the corporate world in the 1980s. In factories you had the traditional problem of soldiering - everyone working at the same pace, often little above the slowest - while in offices the water-cooler became a synecdoche for a culture of work avoidance through the rituals of sociability. WFH offers the potential of fewer distractions and easier monitoring, but also of more granular worker "tuning". The hunt for a better work-life balance is not an initiative of labour but rather capital's continuing determination to raise productivity through greater exploitation.

I suspect that the enthusiasm for working from home will deflate once the pandemic is over. This is not to suggest that the secular trend towards remote and flexible working, with its emphasis on moderation and balance, won't continue, but that we will increasingly see the emergence of different classes of WFH: the professional who manages their own time and goes into a pleasant city-based office for a couple of days of socialising versus the digital peon in a small town whose piece-rate work is closely monitored. Central London will not empty; Mansfield will not be regenerated. Perhaps the most worrying aspect is that those whitecollar roles where trade unions have managed to retain a presence, notably in the public sector, will be increasingly vulnerable to this steady change. Though the technology offers as much opportunity to labour organisation as threat, the reality may be that WFH is combined with a move to greater freelancing and all the disempowerment that entails. Rishi Sunak isn't pushing a conservative narrative in which offices are temples of capital but imagining a world in which atomised labour has to be congregated for indoctrination in a corporate culture.