My recent column in the Times on robots in agriculture:
If you will forgive the outburst of alliteration, the harvesting of a “hands-free hectare” at Harper Adams University has made headlines all around the world, in the technology press as well as the farming press. A crop of Shropshire barley was sown, fertilised, sprayed and harvested by robot tractors, drones and a robot combine harvester, without a human being setting foot in the field.
The yield was low and the cost was high, but the point was made. The mechanisation of agriculture is progressing rapidly towards the point that some crops can be grown with almost no labour. In one sense this is merely the culmination of a trend that began with oxen pulling ploughs instead of people wielding hoes, then continued through the invention of threshing machines, tractors and combine harvesters.
Modern, GPS-guided combines can already cut a field without the driver touching the steering wheel for much of the time, and can cut and thresh a quantity of grain in a day — enough to make half a million loaves of bread — that would have required the work of a thousand peasants in the 18th century.
Harper Adams, an internationally successful technical university, is teaching robots to recognise weeds so they can be hit with tiny spots of herbicide instead of indiscriminate sprays, or better still be zapped with lasers. This could save money, energy and wildlife.
Professor Simon Blackmore, head of engineering, argues that an even bigger benefit could be in providing fleets of small, light robots, perhaps aided by drones, to replace vast and heavy tractors that plough the land mainly to undo the damage done by soil compaction caused by . . . vast and heavy tractors. Robot tractors could be smaller mainly because they don’t have to justify the wages of a driver, and they could work all night, and in co-ordinated platoons.
Watch a video online of celery picking and you will see just how these two worlds combine. A celery-picking machine is something the size of a small house that a tractor pulls very slowly through a field. It takes celery plants out of the ground automatically but passes them up a conveyor to a row of people whose job is to trim and bag them by hand. In comparison with what happens in most factories, it is crying out for full automation.
In the case of fruit and vegetable picking, even though the labour may be “cheap” by the standards of most jobs, the cost of employing it is high, because of low productivity. Moreover, at least 20 per cent of the crop, sometimes much more, does not meet supermarket standards because it is too small, misshapen or blemished. A robot could selectively harvest only vegetables and fruit of the right quality.
It is not inevitable that robots will replace people in any particular kind of farming. Some areas will surely change quicker than others. It will only happen if it pays, if it simplifies rather than complicates the farmer’s life, and if the robots are easy to manage. For example, employing a drone to herd sheep will only happen if a drone’s electricity costs less than a dog’s food and the drone is easier to operate and train than the dog. People will still be needed to plan, decide, and work with the robots.
Just how much has access to cheap labour held back the automation of British farming (and a few other sectors)? History is fairly clear on this point: that if labour is cheap, people do not invest in labour-saving inventions. Slaves were cheap in ancient Rome, so it was not until centuries later that the draught horse, and the crucial horse collar, were invented. Likewise in Japan in the 18th century, there was an “industrious” rather than industrial revolution, fuelled by rapid population growth, in which draught animals and machines such as water mills were abandoned in favour of labour-intensive alternatives. Rice was threshed and milled by men on treadles. By contrast, in 18th-century Newcastle the price of energy was low because of local coal mining, and the price of labour was sky-high, compared with Paris, London or Beijing, according to calculations by the historian Professor Robert Allen. There was every incentive to find ways of getting coal to do the work of people, and mechanisation thrived.
An industrious revolution is a dead-end in terms of human welfare, giving a country an economy with low productivity and so a low consumer surplus to support other jobs, entrenching poverty. Is it even possible that access to cheap labour from eastern Europe, to run the farming and catering industries, partly explains Britain’s sluggish productivity growth in recent years? Together with the subsidising of low-productivity jobs through in-work benefits, we have been pursuing somewhat “industrious policies” — with good results in terms of record low unemployment rates, but perhaps at the expense of automation. Forcing up minimum wages does deprive poor people and students of employment opportunities, but it is also likely that it stimulates automation and the creation of better jobs elsewhere.
Though every job lost is painful for someone, robot tractors are almost certainly going to create more jobs in the long run than they destroy. For a start, there are just not that many jobs left in farming. In medieval Britain farming must have once accounted for 80 per cent of employment, perhaps more. Now it is less than 1 per cent. Automating agriculture further will not throw huge numbers of people out of work. Indeed, if it makes British farming more competitive, it will create more jobs in the processing and marketing industries, as well as in the rest of the economy where the profits of farming are spent: Range Rover dealerships as well as robot dealerships.