Contents
Many basic ideas in economics about wealth creation still trace back to Adam Smith (link in Russian) and his ‘ultimate textbook’ – the foundational work An Inquiry into the Nature and Causes of the Wealth of Nations, published in 1776. On the one hand, this is quite natural. People still work; companies still invest in fixed capital, and natural resources have hardly become less important. On the other hand, the spread of digital platforms, the rise of data and their algorithmic processing as key resources may well shake the very foundations on which Smith built his ideas.
Smith’s legacy has been a subject of regular historical and philosophical analysis. Some researchers see him as a tough economist, and a forerunner of Karl Marx; others see him as a moralist for whom a market without ethics was unthinkable (1, 2, 3, 4). A third camp focuses on analysis of his theory of the stages through which societies pass – hunting and gathering, cattle breeding, agriculture, and commerce (1, 2, 3, 4). A fourth group views his The Wealth of Nations as a textbook on economic history.
For all their differences, however, these approaches have something in common: they all treat Smith as no more than a historical figure. His theory is not used as a working tool to analyse economic phenomena that emerged after the 18th century. No one has tested whether the theory still holds in the 21st century. The foundational work that became the bedrock of modern economics was written by a man who knew nothing of the internet or digital technologies. Yet their spread extends precisely to the pillars that Smith built his analysis on: the nature of labour, the structure of capital and the mechanics of market exchange.
The nature of labour has changed: cognitive operations, which once required human judgement, are now being handed over to algorithmic systems. The structure and model of capital have also changed: data function as a special kind of capital, one that is not consumed as it is used and does not follow the law of diminishing returns. The mechanism of market coordination has changed: the digital platform is an organised marketplace in which rules are set and enforced by algorithmic code.
We test (link in Russian) how three of Smith’s key concepts – the division of labour, the nature of capital, and the mechanism of market coordination – withstand the impact of these new realities. This approach is dictated by Smith’s own method: the economist insisted that theory must be tested against facts and that those facts be real – otherwise they cannot help us understand how undesirable outcomes can be avoided or desirable outcomes brought about.
The digital economy that Smith never knew
Earlier works by researchers, drawing on data from the US Bureau of Economic Analysis (BEA) and the Bureau of Labor Statistics (BLS) across 63 US industries, show that between 2000 and 2015, productivity in digital sectors grew 2.7% a year against 0.7 % in physical sectors. Over the same period, employment in digital sectors went up 29% compared with 20% in physical sectors. At the same time, digital industries – accounting for about a quarter of private employment and nearly a third of private GDP – made up 70% of all private investment in information technology. 15 industries are classified as digital – those whose core product can be delivered in digital form and over the internet, while the remaining 48 are physical.
Our analysis of updated data from the same sources, covering 1997 to 2023, confirms the trend: the economy is switching from investment in tangible IT assets to investment in intangible assets.
Seven of the ten highest growing industries are digital, and occupy the top two places: data processing and internet publishing in the period under review grew almost 11% a year. Computer systems and services added almost 8%. Other digital industries – computers and electronics, broadcasting and communications, rental and leasing, and insurance – grew by close to 4%.
Overall, the digital sector, which comprises three times fewer industries than the physical sector, delivers aggregate output growth comparable to that of the entire physical sector. By every key metric – the contribution of labour, capital, intermediate inputs, and total factor productivity – it outperforms the physical sector.
Where Smith’s theory diverges from today’s reality
Not just quantitative growth lies behind these figures, but also a structural shift: the factors of production are changing their properties, as are the ways in which they combine.
1. The division of labour is turning into a division of access
Smith explained productivity growth through the division of labour. He drew on the example of a pin factory to show that breaking production down into separate operations could multiply output many times over. This happens thanks to increasing dexterity of workers, the time saved in switching between tasks, and new machines that ease labour. In all three cases, the human being was central as the subject of labour.
Digital platforms upend this logic. Labour had long ago ceased to be merely manual effort as a production factor, and predominantly became cognitive activity. But in the past 10 to 15 years, even mental labour is increasingly being complemented or replaced by algorithms (link in Russian).
When looking at a new institution, an economist has to decide whether the new phenomenon is yet another production factor and can be incorporated into the production function. That is precisely how information was dealt with in the mid-20th century – efforts were made to build it into models as a separate variable, independent of labour and capital.
Has information become a separate factor? It certainly has. But it quickly emerged that information changes the quality of labour, affects return on capital, and reduces uncertainty in decision-making. This is clearly evident in the most accessible information generator today: various artificial intelligence models.
In other words, information cannot simply be added to the labour, land and capital triad as another separate element since it redefines the relationships between them. Adding a new factor without a rethink of these relationships merely complicated economic theory, but did not help a better understanding of the way new institutions operate. One of the most urgent tasks of modern economics, therefore, is to understand how to account for these changes.
Taking this discussion further, platforms pose the same question as information – but with greater urgency. Data, technologies and infrastructure do more than complement the triad – they change the very nature of capital. This happens not through the emergence of new factors, but through a new mechanism to combine them: algorithmic coordination of transactions.
It is this logic – the logic of the division of labour in which coordination becomes a distinct function – that platforms are transforming. Smith’s worker specialised in a single operation, such as drawing wire or sharpening pins. A platform, whether a marketplace, a ride-hailing app, a freelance marketplace or similar, is all about access: it does not produce the goods itself, but brings buyers and sellers together, verifies counterparty trustworthiness, processes payments and coordinates logistics.
Coordination, which was once distributed among participants in transactions but is now provided by a platform, has become a separate industry and a separate source of both productivity and rent, i.e. a source of market wealth. There has been a shift from division of labour to division of access.
2. Data are turning into capital that is neither spent nor depleted
Smith defined capital as part of a stock from which income is expected, and divided it into fixed and circulating capital. He also emphasised the distinction (1, 2) between money as a medium of exchange and wealth per se – land, houses and all kinds of goods.
In the modern economy, data do not fit this picture. Data are a special kind of capital. They are capital unconsumed when in use and are non-rival: the same dataset can be used simultaneously by multiple participants without detriment to its value. Moreover, in many cases it grows through network effects.
Still, data per se are not yet capital. They can only become capital when they join processing technologies and access infrastructure.
This is not a unique case: land also only becomes capital when transport, energy and other infrastructure are in place. The same applies to data, which are useless without data centres, processing algorithms and transmission channels. This is the general rule: a resource becomes a factor of production only when there is infrastructure that enables its use.
3. The rules of market coordination are created by the platform itself
Digital platforms not only change the structure of production, but also the very mechanics of market exchange. Smith described the market as a space where buyers and sellers meet, and where institutions – the state and customs in his terms – protect participants’ rights. But the ability to make a deal is constrained by three types of costs. The first barrier is information: how far a participant is able to see a potential counterparty and their offer. The second is transport accessibility: beyond a certain distance, a transaction becomes economically unviable. The third is trust: can a partner’s trustworthiness be verified without turning to authorities, which is costly and slow?
In the traditional economy, people mostly trade with those they know (e.g. trust dominates), the information horizon is narrow, and transport is available only over limited distances. Smith took this picture as a given and did not ask whether it could be changed. But it has.
The information barrier can be overcome through modern search tools, and the totality of offers can be seen. Platform-embedded logistics extend transport accessibility. Reputation mechanisms, such as ratings, reviews, transaction histories, bolster trust and lower the cost of counterparty verification, even if they cannot entirely eliminate uncertainty. The platform does not create a new market space – it just exploits one that already existed but was previously inaccessible.
It is important to understand that this is not about e-commerce as such. From the Middle Ages onwards, markets did essentially the same thing by bringing sellers to one place, reducing transport costs and enabling price comparison and price competition. But while a market was constrained by geography and time, a platform is not. It is also marked by a different scale of data, personalisation and instant counterparty checks. Trade remains an illustrative example here, but similar processes are unfolding in any area with dispersed participants, incomplete information and a need for trust: labour markets, financial services, education, healthcare, etc.
Smith described how productivity grows within production. Platforms show how the productivity of exchange grows – through lower costs of search, trust and coordination. Coordination becomes an independent source of wealth.
In classical economic theory, any scaling up eventually entails monopolisation: growth in firm size leads to market concentration, anticompetitive practices and, ultimately, monopoly pricing. Yet the digital platform has a different logic: rather than just capturing a market, it creates a unified interaction environment for – a digital flow of data for participants (users, services, devices) to exchange information without grabbing market power.
What the classics would say
The ‘gaps’ between Smith’s ideas and contemporary reality raise a question: which parts of the theory are now up for revision? To answer that, we need to look into the structure of economics itself. Since the early 20th century, it has followed a pattern: a hard “core” of basic models, surrounded by a “belt” of refinements and exceptions. The belt has grown, but the core has remained untouched (1, 2). If each new phenomenon demands further complication in the ‘belt’ of explanatory hypotheses, it is time to revisit the foundations.
Institutional theory (link in Russian), behavioural economics and other fields has been patching up theory, but has left the foundations unchanged. But algorithmic coordination, non-consumable data and the new nature of labour are no longer special cases – they are a challenge to basic categories.
It is telling that leading contemporary economists and Nobel prize winners who have studied growth and innovation at various times – Paul Romer, Philippe Aghion and Peter Howitt, and Joel Mokyr – have chosen not to return to the core in the sense we use it here. Without exaggeration, their discoveries have changed economics. Romer showed that knowledge and technology are not an external ‘gift’ but a product of the economic system itself, and incorporated them in the growth model. Aghion and Howitt described how innovations do not merely add something new, but replace the old, and this “creative destruction” is not an external shock but an endogenous, internal, property of growth. Mokyr proved that technologies do not work on their own: they need a cultural and institutional environment. His analysis of the Industrial Revolution and subsequent waves of growth become a classic work. These economists are not interpreters of Smith or historians of economic thought. They are the researchers who built models that work in the conditions of a digital world and innovation accumulation, and succeeded in explaining the mechanisms of growth in this environment.
Yet, for all the great theoretical significance of their works, they do not ask whether the basic categories themselves – labour, capital and land – have changed. Their contribution lies in introducing new variables – knowledge, innovation, institutions – and showing how they alter returns on the classical factors.
To understand the difference between their approach and ours, we must discuss the structure of economics itself. Since the second half of the 20th century, it has developed along the logic described by Imre Lakatos and Thomas Kuhn, who worked separately, but wrote on similar subjects. Lakatos proposed (link in Russian) a model of the research programme consisting of a hard core – the basic propositions that are deemed irrefutable – and a protective belt of auxiliary hypotheses – which are modified and expanded, and thus take the brunt of empirical anomalies. Kuhn described (link in Russian) the mechanism of paradigm change: once there are too many anomalies and the protective belt fails, a scientific revolution occurs involving a revision of the core.
In our approach, the labour, land and capital triad does not acquire another factor. Instead, we go back to the foundations and test whether they still hold. To understand whether Smith’s core withstands the test of new realities, we trace the evolution of the factors of production from Smith to Joseph Schumpeter, the Austrian economist and author of the concept of creative destruction, and Ronald Coase, the British economist and Nobel prize winner. This effort is not a historical survey, but a test of how their theories work in the platform economy, and what they themselves add to Smith’s core.
In today’s world, Smith would first of all notice the transformation we have mentioned: the division of labour becoming a division of access. This is because the coordination of transactions (search for counterparties, trustworthiness verification and concluding a contract) has become a separate operation performed by the platform.
Karl Marx would then take this logic further and ask: who is the owner of this coordination? Data are created by users, but appropriated by the platform; the income from data ownership is algorithmic rent. The distribution of this income is not a technical, but a social question.
At the next step, Alfred Marshall – the English economist and one of the founders of the neoclassical school in economic science – would ask about a consumer surplus ((link in Russian, in its classical definition, it is the difference between the market price and the price a consumer is willing to pay), that is, how the platform affects the welfare of its users. A platform not only creates a consumer surplus through price competition, but also through personalisation of offerings. Lower prices relative to traditional retail, a broader range and time savings can all be measured. They need to be incorporated into welfare theory.
He would be followed by Arthur Pigou, the English neoclassical economist and Marshall’s disciple, who would draw attention to network effects. The value of a platform to each participant grows as the number of other participants increases. This increment is not created by the platform owner, but by the users themselves, but it is the platform that appropriates it. This is then a classic case of divergence between a private and a social product, which calls for correction. Pigou would be able to adjust Smith’s framework through a competitive regime, a low barrier of entry onto the platform and tax arrangements, ensuring that part of the private product comes back to those who create it.
Schumpeter would then see confirmation of his theory of creative destruction (link in Russian). There is however an important caveat: his theory of economic development is grounded in the entrepreneur as an individual champion of initiative. Yet, in the platform economy, the coordination function is increasingly performed by an algorithm. This means that creative destruction in the platform economy is not so much the result of an individual entrepreneur’s will, but a property of the infrastructure itself.
Finally, Ronald Coase would close the circuit by bringing it back to Smith, but at a new level. Coase showed that the market and the firm are two alternative ways of organising resources, and that the choice between them depends on the costs of finding partners and monitoring transactions. The platform represents a third way: it does not reduce those costs with the help of hierarchy, but through algorithms.
But a paradox arises here: as costs fall, the market does not become less dispersed. On the contrary, it is monopolised by a few giants. Coase would explain this by the fact that data – the key resource of platforms – lack clear ownership rights, and he considered property rights as a precondition for efficient resource allocation. But a market cannot efficiently allocate something that has undefined boundaries.
That is, for platforms to work for the common good, rather than for power concentration, the rights to data need to be specified. Only then can data become full-fledged capital – not alone, but in combination with technologies and infrastructure.
This brings us back to Smith, who understood that the market needs rules. Coase would probably be more specific: in the platform age, those rules must be about data.
The chain we have traced from Smith to Coase is not a gallery of theories, but a sequential application of the method of returning to the core. Data ownership, a platform consumer surplus and network effects as an externality, continuous update as a property of infrastructure, and the specification of rights as a condition for efficient allocation are all dimensions that do not disprove Smith’s core, but rather incorporate it in a more complex system of relationships.
Labour, land and capital retain their place in the foundation. But their combination is now mediated by (1) coordination as an independent function, (2) data as a non-consumable resource, (3) processing technologies as a source of welfare, and (4) infrastructure as the environment that sets interaction rules for all elements. This is not an expansion of the list of factors, but an evolution of the economic system itself – the type of evolution that Smith’s method of testing theory against facts required to be documented. It is precisely what Romer, Aghion, Howitt and Mokyr are describing, each from his own angle.
Our analysis does not close the debate on Smith’s ideas, and leaves several questions that require further research:
- How do we determine whether a platform is a market or a new form of firm from a regulatory perspective? The answer will determine the design of antitrust oversight.
- Can algorithmic rent be measured and taxed without weakening companies’ incentives to innovate?
- How should rights to personal, big and public data be delineated? Where is the boundary between the interests of platforms, users and the public at large?
Adam Smith taught us to refrain from final answers, but ask new questions. Today, 250 years later, that approach remains as valuable as the ideas he left behind.
The full text of this research is published in Voprosy Ekonomiki, 2026, No. 7.
