Myths of Value: Why the IT Revolution Never Lived Up to Its Economic Potential
It gives great, if rather quicksilver, insight into the tenor of our times that the statement “The IT Revolution has failed to produce meaningful economic benefits,” seems not only absurd but self-evidently absurd. The great totems of the digital millennium are not just universally recognised for their valuation, for their notoriety; they are so vastly influential that they have become the period itself.
From schools of economic thought to dramatic traditions, our times (whatever and whenever they may be) are primarily defined by the stories we tell ourselves about ourselves, stories that are no more beholden to fact than they are insensible to them. Narrative, after all, is a great filtering device that aids us in our compulsive attempts to reduce the world’s incomprehensible mass of complex interweaving developments into simple truths. Our particular insistence on the idea that we live in a time of unprecedented tech-driven economic advance is, therefore, primarily based on a story told to that effect — stories that some would place under a less-than-absolutely-charitable blacklight, and call ‘lies’.
According to the economist Robert Gordon, and his 2016 book The Rise and Fall of American Growth, we in the 21st century remain beholden to expectations of progress — of its inevitability; of its implacable goodness, its rightness, its beneficence — that were developed during those times of truly unprecedented technological advance that episodically defined the period between 1870 and 1970; expectations of progress that, so Gordon would have it, are no longer being substantiated by output. Productivity rates have slowed, consumer welfare is in decline, wealth and value are not being created in the abundance we might have expected. We may be cash rich in our Early Digital time, but by most criteria able to meter the degree to which a society lives in consummate civility — criteria both economic (real wage increase relative to the price of goods) and more abstract (the number of words in our language) — we are becoming increasingly deficient, the promise of a new iPhone every year be damned.
How could we have made such a misinterpretation of where we are and the shape we’re in? Perhaps through a misunderstanding of that most contingent of terms — ‘value’. In the Early Digital age (the period marked by the beginning of the Third Industrial/IT Revolution, still ongoing) ‘value’ is not so much a label or an honorific term or a tangible outcome, but a metabolic process, a speculative quantity that mainly consists of the appearance of value being created. Partly because we live in worlds characterised by digital abstractions (and the economic abstractions of speculation), ‘value’ is no longer something that is considered intrinsic. It is no longer something, at least where ‘value’ is concerned with technology, that rests in a fabricated, useful object, to be appraised in terms of the object’s ends.
In other words, this revolution, our revolution that has transformed and expanded our worklife and upended our social habits has not generated the value expected, and with value its fiduciary compact, wealth. It has behaved overloyally to the literal definition of the word ‘revolution’: it has spun in circles, merely redistributing and re-concentrating wealth that ostensibly already exists.
If all this is true, how could such astonishing feats of technological progress fail to yield the transformative effects they seem so plainly capable of? What obstacles stand in their way? And how, presuming the possibility, can they be overcome?
1870–1970: The Level-Up Century
Robert Gordon’s assessment of the interaction between technological progress and what we would term an adequate ‘progression’ in living standards reveals just how miraculous those advances of the century from 1870 were, particularly within the 1870–1940 period.
That 70-year period saw the unheralded procession of major inventions that formed the Second Industrial Revolution. The internal combustion engine, the rotary kiln, the electrical generator, the telephone, the light-bulb, the automobile, the electrical grid, the cinematograph, the vacuum cleaner, cellophane, silicon — all of these things formative to the fabric of modern life and work came newly into being.
As Gordon has it, the changes made to the quality of work by these new technologies between 1870 and 1940 were unalloyed goods. It’s hard to disagree with him. Work became safer, roles of employment became broader and more interesting, and real wages increased steadily. Not only did work pay better than it ever had — workers needed to work fewer hours-per-week than they’d had to before. Machine-abetted productivity invited hours-worked-per-week to fall, and let the weekend as we know it come into being. Appreciable amounts of leisure time became the preserve of more than just the ‘leisure class’.
This period saw certain work-related practices that we now rightly adjudge as incompatible with civilisation — things like child labour — becoming extinct via their becoming unnecessary from a practical perspective. It now made not only humanitarian sense, but economic sense too to keep children in school, to produce a more technically competent workforce (though not without controversy; believe it or not, the idea of universal literacy had its opponents). More and more women began entering the workforce, though not nearly as many (and not nearly as many of those who were unmarried) as would enter it during World War II.
Consistent unionisation, epitomised in America by the heroic likes of Samuel Gompers and Walter Reuther, helped exert the requisite counter-pressure that would drive productivity up. Unionisation is often invoked as a natural handicap towards both productivity and service quality — and when the power of unions is harnessed to corrupt or shiftless ends, that is often the outcome. However, as was the case during the early years of the Franklin Roosevelt administration, unionisation expanded the consumer base by endowing workers with the bargaining power required to maintain wage growth.
Gordon assesses two other benevolent behemoths — credit and insurance — that helped stimulate growth via opening new channels to consumption. Loans against life insurance were frequently used as a source for a down payment on houses and cars. These private concerns were backed institutionally and meant that the average person’s quality of life increased in more-or-less direct proportion to productivity increases. Innovation was doing the job, and its effects were not only felt by those occupying society’s lowest-hanging economic basket; those people arguably felt its benefits most keenly.
Moving through the middle of the 20th century, the post-World War II period produced another concerted wave of world-changing innovations, from polyester to transistors to the video tape, hard disk, laser, and the personal computer. The Level-Up century was capped off by the extraordinary exertions of ARPA, which foregrounded the coming of the internet.
1970–2020: The Fall-Off
Technological development does not seem to have tapered off by comparison of 1870–1970 with 1970–2020 — indeed, judging purely by productisation and the wide availability of technological consumables to a mass market, such development has accelerated to the precipice of actuarial escape velocity. And yet Robert M. Solow’s observation in the middle of the Third Industrial Revolution tempts us to doubt: “We can see the computer age everywhere but in the productivity statistics”.
The slowdown in conditional-improvement-relative-to-technological-development in 2021 seems all the more marked given that astonishing pace, regularity and drama with which technological developments themselves seem to come along. In the uplands of the 1940s, it was forecast (and logically, given the progress of the prior 70 years) that within decades automations would bring humanity’s necessity of labour down to an absolute minimum, and work hours to a negligible level, putting the majority in the leisure class bracket and thus inaugurating an entirely new chapter in the history of society. Instead we have seen the automations promised lead to greater prolatism: conditions wherein we work longer hours, not out of practical necessity, but out of a perverse sense of work’s moral good (a ‘moral good’ that is, in torturous ill-logic, simply a confused conflation of the acquisitive instinct and the physics of envy with moral good).
This strange turn away from most quantifiable notions of progress seems to have coincided with a change in temperament. The post-1970 period is the first time we see the beginning of the kind of financialist utilitarianism (productivity in order to reconcentrate wealth) which has come to dominate the way in which we develop and prioritise mass market technology. We see it early on in a shift to the priority of American healthcare around 1970. As Gordon states:
“Since 1970, the health care system has shifted to more expensive, capital intensive treatments primarily provided in hospitals that have led to an inexorable growth in medical care’s share of GNP increases that most scholars agree exceed any improvements in health outcomes.”
‘Well, so what?’ you might exclaim. ‘GDP is still going up — a bit of deregulation didn’t do anything to halt that!’ Well, yes, this is true — but for the most extreme of circumstances, GDP always grows at least a little. Rather, it is the change in the rate of progression that is telling. Robert Gordon uses TFP, ‘total factor productivity’, as his main north star metric when it comes to making a broad survey of the economy throughout the time period he surveys. The way in which TFP growth has changed is instructive as to the way in which our Third Industrial Revolution of social media and gig economics has floundered compared to its predecessors.
By comparison of these metrics Gordon concludes the Third Industrial Revolution’s greatest TFP bump occurred during the paper-to-digital switch of the decade between 1994 and 2004, and then slumped. But for the COVID-enforced rise that followed the migration to remote work, the slump has been little disturbed since.
The GDP Problem
There has been, in those recent seasons of decline, a coincident rising vogue for measuring economic growth with respect to GDP-per-capita, as though this statistic were what a first impression suggests it should be — the average rate of productivity of an individual, as that rate commutes to an individual’s share of overall production that the individual themselves benefits from. However, GDP-per-capita is meaningless when used in this way. The meaningful metric here is consumer surplus; if it increases, consumer welfare grows. It is a felt improvement in conditions, one that by nature reaches and affects consumers, not an abstract one like GDP.
Authentic improvements to ‘consumer welfare’ are brought on by new products, increased product quality, and proportionately lowered prices due not to savings made on the supply chain (e.g. by outsourcing of labour) but to the surge of total-factor productivity (TFP). TFP surges of this kind are provoked by innovation. This conflation of innovation and eventual consumer benefit has resulted in an unfortunate syllogism — that anything that causes prices to drop is ‘innovation.’ However, a great many of the price-dropping pillars of our Early Digital commerce — Amazon, Uber — come not through the creation of value, nor really through any meaningful technological innovation (nothing comparable to the universal transformative power of, say, urbanisation or the combustion engine), but through a creative approach to the circumvention of legal codes. Price drops brought about by these companies have come about via the outsourcing of risk, the slashing of the overhead and the imprecariation of the worker (which leads directly to lowered wages), creating a zero-sum dynamic to consumer welfare. Under these auspices if the consumer benefits, because no discrete value is being created elsewhere, then someone somewhere else on the supply chain is shouldering a corresponding deficit.
We can see in prior epochs that the zero-sum dynamic is not necessary to procure better outcomes and lower prices for consumers — in the Level-Up century, consumer welfare rose because innovation had caused productivity to rise and prices to drop, with healthy unionisation in force to ensure competitive wages and keep spending power high. Now, imprecariation has caused an artificial price decrease, bolstered by a certain amount of the most notional productivity increases brought about by a culture of prolatism, which itself has been enabled by the erosion of worker’s rights (no overtime, an embattled resistance to the provision of sick pay etc.)
One of the reasons GDP growth rates cannot possibly be used a representative proxy for consumer welfare is because the “improved working conditions, the end of backbreaking toil, especially regarding labour at home, and increased leisure time” which indubitably contribute to welfare, and which have been immeasurably enabled by innovations during that key period of 1870–1970, are not encompassed within growth rates.
Even if we believe the GDP gospel, ostensibly composed in order to give credibility to the idea that what is productive for GDP is broadly beneficial for society, the innovations of the Early Digital period (taken from about 1970) don’t live up to the billing. They contributed only 7% of American GDP in 2016, and increased no higher than 12.9% in 2020 even after the global switch to remote working. There is a plausible argument, one not much assessed by Gordon that the innovations of the Third Industrial Revolution have enabled some productivity rises in other industries, and so cannot be understood in their entirety within a discrete category of their own. Any such allegation would have in any case to stand against the counter-argument that, in many notable instances, these technologies have actually reduced the value of product (and workers’ welfare and quality of life with it).
For example, it is axiomatic, a running joke even, the way in which technology wiped billions of dollars off the value of the music industry, reducing its revenues from a $21.5 billion late-90s peak to under $7 billion in 2015, without any kind of measurable redistribution of that value elsewhere, except perhaps to the luxury fashion and cosmetics industries for which the music business is now essentially a bespangled prop.
The loss of productivity and collective gains are not a Great Recession thing, either, nor will we be able to blame the coming decade’s presumable continuation of our observed slump on the COVID economy. Gordon’s historiography points to the Great Depression as being the venue of one of the great leaps in innovation and productivity in that hallowed century between the two ’70s. He insists that the New Deal (and especially NIRA and the Wagner Act) promoted unionisation and “directly and indirectly contributed to the sharp rise in real wages and a shrinkage in average weekly hours.” This subsequently “produced substitution of labour for capital, capital deepening and boosting productivity.” There is no counterfactual consideration in Gordon’s work to consider whether or not those leaps would have taken place without the Depression; even still, it does prove that great innovation/productivity gains can occur during, or even in spite of, global downturns.
However, the reason why those gains may have occurred during the Depression in itself gives us fascinating window into certain deficiencies of our time. Given how pressurised the US’ market economy was when the Depression hit, “the competitive pressure…produced incentives for efficiency and improvement of production, irrespective of whether a sharp drop in output had occurred. Thus, it seems that the TFP surge during the 1930s would have occurred even without the Great Depression.”
The conditions within the Third Industrial Revolution, meanwhile, have prized both vertical and horizontal integration, management culture, and the building of trusts — one of its most prominent architects, Peter Thiel, is an avowed champion of monopolies, and the erection of digital trusts has formed one of the most profound ongoing controversies of the Early Digital period. Gordon himself, for that matter, does not seem to be any particular apostle of competition. He certainly doesn’t remark on possible links between his posited lull in value-creation since 1970, the evisceration of the trust-taming Federal Trade Commission in the 1980s, and increasingly aggressive monopolistic behaviour by large tech companies in the decades after.
All these numbers and ideas make for an interesting wound to the collective sense of self. The Early Digital self-conception is that we are richer, more advanced and more progressive than our forebears — not just richer, but better (and the common colloquial confusion of the two terms as synonymous might hardly be understated when pondering how we got to this point). Even if we are more challenged in matters social and ecological, we have technology. But, if that technology is failing to have the kind of impact which generates tangible outcomes for consumers and workers, if we re-identify that consumer technological pedigree of which we are so proud as primarily a pack of luxury trappings, and take it away, what are we left with?
Stratified income inequality without evident opportunity for reversal, lowering levels of education, lowering happiness. Very little R&D investment that would give flight to genuinely paradigm-busting ideas. A culture, and with that culture a collective life, free of the vision and nerve required for progress.
Why do you need nerve and vision? Because they temper and condition productive approach to risk. And risk is something we cannot do without.
Why There Can Be No Value Without Risk
Now, to the pith of a bete-noire:
“We wanted flying cars, instead we got 140 characters.”
…so said Peter Thiel.
How can this course be corrected? Changes to the way we make narrative hay from economics might be a start. Economics is the literature of numbers, more art than science — that many people, including fledgling economists, believe otherwise is testament to the power of narrative that we are just now about to discuss — and is governed in part by the principle of speculation. This means that things come to have an inflated or deflated value based on the degree to which they are perceived to be valuable. Things that come to be perceived as valuable are the subject of investment — financial investment if the ‘things’ in question are tradable assets, intellectual investment if the ‘things’ in question are abstract quantities like ideas or stories.
If you doubt this is so, consider the way in which headlines have been thick with mentions of NFTs, the most contingent-in-value investments imaginable, in recent months; or the fact that, while Elon Musk is now so popular that his mere Tweet-things can cause vast reversals of fortune for crypto investors, it is only by means of a Ponzi scheme (and rampant betrayal of the company’s own climate ethos) that his Tesla can give the world the misleading impression that it makes money. Nevertheless, half a billion’s worth of NFTs have been sold, and Tesla shares are just getting more popular (especially with the growing cohorts of market-hype-susceptible retail investors). These are the stories that people prefer.
Much popular understanding around the concept of broader value is tied up in the shorthand of ‘GDP’, and its extension ‘GDP-per-capita’ — common understanding rests with the notion that what is good for GDP is good for an economy, and that what is good for an economy is good on balance for each contributing member of that economy. This is taken as given in legion public sector whitepapers, financialist magazines, general interest newspapers. If instead this sense of economic good was centred around something so innocuous as the phrase “consumer welfare”, then principles of nudge theory would suggest that some long-term shift in priority, in our conception of the ends and purpose of an economy, could occur.
Revising the terms we use as a shorthand for economic progression might have any number of positive effects in determining collective priorities when it comes to innovation. A second powerfully influential factor on the future of innovation concerns R&D spend, and where it is allocated. R&D had a restricted, regionalised effect in its early decades at the turn of the 20th century, often used to curry favour for businesses threatened with antitrust litigation, but rapidly grew into the chief driver of invention. It is now most concentrated in manufacturing, a declining economy sector — though, if you have used an iPhone or a GPS in the recent past, you have harnessed the products of research begun (and technical risk shouldered) by the American private sector, and have thus at least some practical feel for the value of government research money.
The reduction in innovation and the way in which it is has helped to entrench the inequalities of our current market model also constitute an argument in favour of more robust antitrust policies. We have observed the way in which innovation ultimately comes to drive up consumer welfare; and we are well-acquainted with conventional understandings of the relationship between competition and innovation, as one spurs on the other. It is one of the black spots of Gordon’s book, the lack of appreciation of the role institutions (and the aggression of their antitrust policies) play in determining patterns of investment and thus the shape of innovation.
One reason why the pace of innovation has slowed is owing to the erection of VC culture and a progressive unwillingness within the governments of wealthy economies to invest heavily in R&D, primarily because agents of those governments do not have the will, the understanding, or a combination thereof to recognise the vast influence governments must play in funding the kind of fundamental science research that eventually gives rise to innovation (even though, as we’ve just seen, the foundations for the ultimate private-sector tech success stories like Apple were laid in part by work done on the public payroll). For those who may retort to the effect that “Venture capital will/could/is providing this impetus with a speed, expertise and willingness to take risks that governments cannot match,” may they be reminded that venture capital is designed to take and exploit market risk, while steering studiously clear of technical risk — in fact, one can hardly build a profile as a successful VC if one’s priority is encouraging technical risk, which promises no returns during an appreciable timeframe. This is why VC culture has encouraged an explosion of economic activity with no profound technical gains. This is why we have an economy that is tailored for the victory of speculators, and in which the work of the big new players is poisonous to real-value investors.
The conservatism of VCs is hardly solely responsible for this state of affairs — they have merely assumed an endemic sensibility. The way in which the pace of innovation-driven productivity increases has slowed so far in the Early Digital period, in a period where tech companies have found it easier to perform horizontal mergers and build monopoly clout than any American companies have in a century, may not be coincidental (nor is it likely coincidental that R&D’s period of least impact during the 20th century was when it was used as an apologist mouthpiece for companies fighting antitrust suits). It is equally unlikely to be down to chance that the period of 2010–20, wherein high profile horizontal mergers and aggressive acquisitions by the ‘horsemen’ increased, was so light in pronounced innovation beyond marginal utilities, despite being the most cash-rich in the history of money.
A final conclusion we can draw from the Gordonic analysis of innovation and productivity over time is that periods of profound advance in both ideas [1] and the productisation of those ideas [2] coincide with periods of diligent regulation presided over by institutions, and high competition within the market. The integrity of market competition in big tech is a contradiction in terms, an oxymoronic joke, as a decade of flagrant herculing has shown. Regulation, likewise, may be on the way to achieving critical mass in the public imagination; there is reason to hope its time is beginning.
If institutional actors can get behind resolutions to better guarantee the free passage of young companies through the proving realms of big tech, without fear of instant acquisition or being unscrupulously plagiarisied out of value, and continue to regulate and dismantle big tech trusts as they emerge, we may at least begin to create some of the base conditions required to begin harnessing the full power possessed by continuing technological innovation to improve consumer welfare, to generate new wealth and hithero unimagined value, not merely recirculate and reconcentrate the old.
[1] …and there are no epochs in which paradigm-shifting ideas have both sprung and then been productised within a single generation. A given subsequent generation’s cohort of doers is generally engaged in the productisation of the ideas thought up by the prior generation’s thinkers, releasing the wealth latent in them, while the thinkers from the same cohort labour on research and studies which surface ideas still 10–30 years (at a 20th century rate) away from actualisation [xx]. For subject-adjacent proof of this, consider the personal-computing and internet-related PARC/ARPA work first entertained in the 1960s and 70s subsequently brought to the realm of product by Microsoft and Apple, whose operating systems still in some ways never harnessed the full potential of their forebears’ vision.
[2] This is not necessarily owing to technological limitations, but just as frequently proceeds from the difficulty which any new ideas face in gaining common purchase quickly, and the generally restrictive and innovation-resistant bureaucracies, public and private, that new ideas must penetrate in order to be actualised.