Why the digital economy is set for a correction

The driving down of prices by digital technology may soon tip the balance of business

I have a grand theory: major parts of the digital economy are out of equilibrium and will go through a correction. Which parts do I have in mind? Without naming any brands: ride-hailing, parcel delivery, food delivery, retail warehouses… Essentially, all the bits where digital meets workers.

These areas all have one thing in common: the workers in them are directed by algorithms. From our perspective at reputable.AI, the algorithms are narrowly optimised on a limited view of efficiency. Efficiency is simply defined as tasks completed per unit of time.

The notion of correction in markets

The idea of corrections is common in financial markets. An asset category, individual stock or currency, trades at a high price or a high multiple for a substantial time and then something changes. The something is political or economic in the case of a currency. In the case of a single firm its prospects change.

Perhaps the firm issues an earnings warning, or goes through a change of management. The asset, which has been trading in a range, or temporary equilibrium, then changes to a new equilibrium position. We saw this dramatically when Britain voted to leave the EU. Sterling, which had been trading at around $1.48:£1.00, moved to a range centred on $1.29:£1.00.

What we are suggesting here is a correction affecting whole sectors of the digital economy, because they are in fragile and temporary equilibrium states.


Equilibria are everywhere – in physical systems, markets, ecological systems, emotional states, markets… If you are stationary, or moving at constant velocity, you are in physical equilibrium. A rugby scrum is in equilibrium if neither side is gaining advantage.

Equilibria can be stable or unstable. Some, like the scrum in rugby, are prone to sudden collapse. This is where we are suggesting that several of the digital-meets-physical players are – the ride hailing and delivery companies. We think their temporary equilibria could fold quickly.

The consequences of cheapness

For a long time in the West we have held the idea that everything we really need or perhaps even want should be as cheap as possible. It’s fine for prestigious things – cars, jewellery, fashion – to be expensive, because the expense is part of the prestige, but not routine things like heating, food, transport, broadband, shoes, medicines…the list grows with affluence.

There’s certainly an expectation that one of the consequences of Digital is that things just get cheaper and consequently, we can consume more of them. But how do these things become cheaper and what are the consequences of falling real prices?

Two things, above all else, have contributed to the decline in real prices for the digital-meets-physical category: taxation and algorithmically driven labour efficiency.

Tax minimisation is facilitated by an international tax regime dating back to the 1920s, when it was reasonable to tax corporations based on physical presence. This doesn’t really work in a world of transfer pricing, where rents can be extracted from subsidiaries in high-tax locations for the use of corporate intangible assets such as brands, patents, and software, thus minimising profits.

Taxation will, eventually, get sorted. France and, surprisingly, the UK seem to be leading the way on this.

Uberisation: maximising labour efficiency

The decline of unit labour inputs is another matter. If we think about the design of, for example, a work management system used in a warehouse, its major purpose is to avoid employee downtime. Complete Task 1; start Task 2. Simple. Nothing new. Henry Ford and Frederick Taylor would recognise this.

What is new, or at least increasingly prevalent, is just how much of the service workforce this now affects: delivery of parcels, delivery of food, delivery of shopping, warehouse work, repair work, call centres, meal preparation (dark kitchens), content moderation. The French have a phrase for it: l’ubérisation.

What are the consequences of this new-found efficiency? And is this really efficient? Let’s look at the second of these questions first.

Exogenous costs

Exogenous is a word favoured by economists. It simply means outside or coming from outside.

All businesses seek to minimise costs and legally have to account for the costs they incur. However, the costs businesses incur are not necessarily the same as the costs business cause. In the cases we’ve been considering, there is considerable debate about whom some of the costs belong to.

Consider accidents or ill-health caused by work. An individual can sue her employer if the employer has failed to provide a safe working environment that results in harm. But what about ‘routine’ injuries at work – mental or physical? The cost of treatment lies with the NHS (society).

It really all depends on where costs are measured. Provided a cost can be legitimately regarded as not a cost of doing business – an exogenous cost – it doesn’t feature on a profit and loss account.

Dominance before profitability

The position is further complicated by deferred profitability. If we consider Amazon and Uber in particular, they have clearly followed a growth model that defers profitability in favour of very wide market dominance. We define wide as both geographic spread and product breadth.

This further complicates any consideration of whether these are genuinely sustainable businesses. We’ve yet to see them operate in a steady state; investors have been funding a race for growth, based on a conviction that above-market returns will at some point follow.

Sustainability and ‘snap back’

At reputable.AI we are largely scientists and engineers, so we tend to think in terms of systems, equilibria, and stability. Our theory, which only time will prove, is that probably all the digital-meets-physical players are in states of precarious equilibrium and that these equilibria will break. Here’s how we see it diagrammatically.

Assuming our theory is true and there is to be a new equilibrium, what would the new equilibrium look like? Our analysis of these digital-meets-physical players differs by category, but low prices for consumers is a common theme. Our view is that prices may well rise in the short-term as returns to labour increase and the costs of regulation are factored into prices.

The red arrow in the second diagram suggests the direction of re-equilibrium. However, a secondary factor of the rise in returns to labour could be an acceleration of automation, for example through increased automation in warehouses and transport. It’s important to realise that this is a dynamic equilibrium.

The implications of equilibria

If you are interested in a long-run historical view of such matters, go to Carl Benedikt Frey’s excellent book The Technology Trap. Frey looks at these balances from the Industrial Revolution to the present day. He also points to the fact that there are 3 million people employed as drivers in the United States and the benefits of automation do not get distributed evenly.

The implications of new equilibria are quite serious for investors. We are beginning to see the implications in share price performance in both specific stocks (no names) and sectorially. There is a quiet revolution underway as fund managers start factoring in a new equilibrium around the environment. Nobody hands Extinction Rebellion a certificate for winning the argument, but the investment industry is alert to this.

As a business, we believe that there are huge upsides to algorithms and AI but there are limits to how much human ‘efficiency’ businesses should seek and we need to be very careful as to how we define efficiency.

We are actively seeking to consult for all the physical-meets-digital players, because they’ve achieved some tremendous advances and some things simply need to be recalibrated.