Automation and commoditisation

Is there an unexpected, dark side to automation?

There has been a great deal of comment on the effect of automation upon employment. Those who - despite historical experience and current productivity data - take a negative view foresee mass unemployment. However, there are a near-infinity of useful tasks to be done that happen not to be monetised, and which needs state funding. If automation created wealth, and the state could lay its hands on that wealth, then the issue of unemployment for low skilled individuals need not arise.

Those last two "ifs" are, though, considerably unsure. Let's take them in order.

Will automation generate wealth?

That question seems to be nonsensical. Automation makes manufacturing and services more efficient, extends their reach, increases their quality. Of course it creates more wealth.

Let's think about that in the context of an entire industry. The figure symbolises an industry before and after automation. The panel on the left shows the "before" picture. Firms have been stacked up in order of their unit cost of production, with their cumulative capacity running along the horizontal axis.

The market size in indicated by a vertical arrow. The marginal producer sets the price, and the industry added value is the area between the price and the individual costs of production, shown in yellow. The industry added value consists of that and the wages embedded in the cost element.

On the right, we find the same industry after automation. Insofar as technologies are similar, costs become very similar. Wages are very considerably eliminated. Automation also encourages the pursuit of scale, so the surviving firms become larger. Market volume of course increases, as shown, and it cuts the cost curve to set the now much lower price. As is evident, industry profit largely disappears. (Individual firms may do quite well, as large volumes times small margins may equal or surpass large margins but smaller volumes.)

Consider, though, what happens in aggregate. GNP is the sum of added value. But here we see that added value all but disappears, at least on these diagrams. Output increases, costs fall, quality improves but the economy does not get larger and employment falls. (This is called "commoditisation", and is a common phenomenon in ageing industries, or industries with many independent producers. Automation, in this sense, is accelerated ageing.) The implication is that whilst output may accelerate and prices may fall, sensu strictu GNP will stagnate. Profitability will be depressed, and without an option for any renewal that can be expected to last. Return on assets will be depressed, chronically, with wide ranging consequences. Amongst these:

Pensions, already a major issue for the ageing West, will become harder to fund.
States will find it inexpensive to accumulate debt, and will do so.
Firms will either drop their hurdle rates for new projects, cumulate cash or quietly liquidate themselves through buy-backs, dividend distributions and the like

Aspects of these consequences are already with us.

Automation can, then, have unexpected outcomes. If it proceeds quickly, and with some universality, it can destroy industry profitability. (Cost cutting in the 1990s had a very similar if less dramatic effect. Firms converged on a benchmarked standard, and industry margins fell: in the case of desk top computers, from well over 50% of sales price in the 1980s to around 4% at the turn of the century. No firm could stand aside from this process, but nobody except the consumer did well from it.)

Can governments extract this wealth?

States have relied on income, speciality and general sales taxes for the bulk of their revenues, with corporation tax yielding relatively little. If employment by companies is to fall and states are expected to subsidise lives, one way or the other, this formula will have to change.

We need to consider three important features of the world ahead.

First, the old rich world will be responsible for a shrinking proportion it world output. It is already at below half of world output, and guesses at 2040 put it at a quarter or a fifth. It will not be in the driving seat, and will not be able to mandate solutions.
Second, automation is no more confined to the developed world than are cellphones or other technologies. The choice of whether to automate may be defined by the activity as much as by general wage levels: you cannot undertake high frequency trading with human traders, however smart. Whole swathes of industry will require an increasing fusion between systems and human staff, with the demands on those staff in creasing in complexity and intensity. By sheer numbers and by demographics, the emerging economies will have more of such people than will the old rich countries.
Third, it is probably wrong to think about the business economics of the future chiefly in terms of countries. At the beginning of the Twentieth century, the nation states of Europe each had a number of car makers whose task it was to service internal demand. Everything from beer to balloons was produced domestically, and the government was sovereign over everything within its borders. All of that had changed by the end of the century. Fewer, now more or less international car manufacturers drew upon a network of hundreds of thousands of suppliers. (Toyota alone had 20,000 just within the US.) A "de-located" cloud of vehicle manufacture spread out across boundaries. It is critical to future economic success that states should attract new initiatives from within these 'clouds'. Th reality is that whilst there is a limited amount that states can do to make themselves differentially attractive, they can very easily achieve the opposite. States that adopt hostile regulations or high taxes will drive away nascent industries. Ill-considered German environmentalism weakened its leading pharmaceutical industry, chiefly by banning aspects of biotechnological research. Innovations, new plant, the concentration of skilled people gradually concentrated elsewhere, increasingly excluding Germany from that particular cloud.
States are also constrained by knowledge that its agencies share with their peers. For example, what central banks know about macro-economics or what health regulators understand about health economics is knowledge which sets profound limits to what a politician can suggest. Independent think tanks and specialised journalism translates policy absurdities into terms that the public will find digestible. This aspect of public life has become known as 'post-democracy'. It is not yet widely understood by the public.

In review: change will be enacted in a global cauldron which nobody controls, in which automation will be a general truism and in which the emerging nations will be fielding more graduate workers than the old rich world has citizens. Nation states can do little to attract new initiatives but a great deal to repel them. (Offers such as tax breaks lead to rapid races to the bottom, as every nation is forced to join in.) Policy makers operate in a tight box, therefore, the edges of which are defined by best practice as shared amongst industries, regulators and state service professionals.

And in all of this, automation may be having the perverse effects with which we began: slow growth, declining profitability and employment. Capacity may be much expanded, but static national incomes imply deflation.

There is a further complication to consider. Cities are now the focus of the world economy. As automation severs the link between population centres and manufacture, what is done in cities will be increasing intangible, carried out by highly educated services workers who earn high salaries. Cities will continue to price out anyone who does not fit into this category, or offer services to them. The gap between the national elites and the non-participant hinterland will grow: a cultural, social, economic and political gap. the typical New Yorker expresses attitudes and ambitions that more closely match those head in London than either of those country's hinterlands. Britain's South finances life in the rest of the country; as does the Isle de France, the Chinese and US coasts, the Tokyo sprawl. Wealth generation is highly focused in every one of these.

The current model of "one nation" welfare is that each country taxes its economic uplands and redistributes this wealth nationally, as services and as welfare, such as pensions. In the stagnant world that automation may bring, where cities belong more to specialised organisational clouds than they do to nation states, how long can that persist? How long can the old rich world behave as it did when it had a monopoly on everything, from military power to technology, culture to economic strength, when it no longer has a hold on any of those? If its citizens plunder their cities, the engine will soon stop turning. If they neglect their hinterland, democracy will soon elect unfortunately leaders. Something has to give.