Cute, huh? Everybody loves a red squirrel. Maybe it’s the effect of Ladybird books or Beatrix Potter (who I believe was secretly a Furry). Despite the irrefutable fact that a squirrel is essentially a rat with an expensive PR agency, you have to love them in all their acorn munching, tree scampering, tasting brilliant in Majorcan style paella cutitude. Because that’s what we do. We love animals that look cute. Trust me, the internets are full of them.

But what matters  in a system? The cute species or the OMG fugly earthworm that aerates the soil and allows the plants to grow upon it, and therefore provides the plant feedstock upon which the cute creatures live? Why doesn’t the WWF use a blind ugly worm as its logo? Naturally, we tend to focus attention on the big cute species, but that’s because of their significance to us not to nature.

And so to industrial policy. The red squirrel effect is an effect in industrial policy that I have just made up to describe the tendency to focus on cute species rather than systems. For good Ol’ squirrel Nutkin, you can insert a given champion/sub-sector/favourite policy example and extol its virtues to your hearts content and then turn round and say that therefore makes a case for some kind of industrial policy or other. It’ll look superficially great. You’ll compete to have an opportunity to associate yourself with its success. Then in twelve months, it’ll launch it’s new product. A Leyland Princess. It’ll tank. There’ll be angry enquiries. Taxpayers, that subset of the population represented by the Daily Mail, will be furious. Reports will be written, recommendations made, policies changed and you’ll be back at square one. That’s how industrial policy worked in the UK in the 1970s, for a given value of worked of course.

The fact of the matter is that if you do want to operate an industrial policy, you have to do so given that you have concern for the system in which industrial activity takes place. If you don’t have an ecosystem niche for a red squirrel then it doesn’t matter how many red squirrels you try to support: they will all die out regardless. If, on the other hand, you have a forest with the right type of basic feedstocks, the right overall balance of species within it, and importantly the right predators, then you have a chance of introducing a red squirrel colony and having it work. (Incidentally, the Heck brothers who, supported by Goring, tried to breed back the extinct aurochs found out the hard way: their breed, Heck cattle, were never able to thrive in the wild because they had not co-evolved to work within that system. As a result, Hecks that escaped into Belarus died out pretty soon. The remainder became a) domesticated and b) studied and found not to be so similar to the original aurochs, to the extent that Spanish fighting bulls are probably a closer match. But that’s for another day.)

So if you are going to have concern for a system, then you need to start with the root species. For any given industrial system these are going to be “general purpose” or “basic and enabling technologies”. Both of these descriptions are actually misnomers. The ‘technology’ is technology in an economic, not engineering sense, and therefore means ‘way of changing inputs to outputs’ and are described as people, methods and things, i.e. capital, labour and way of combining the two to make an output. I emphasise this point because it is not sufficient to have a general purpose technology without the people and know how needed to make it work. The feedstock of industrial systems is therefore a stock of human and technical capital that enables technologies higher up in the food chain to access the root feedstock. Now the existence of root feedstock is something that government should be able to help manage: after all, it has oversight of the bulk of the education system turning out people who understand the feedstock and its access points right? (Er,,, well,,,,)

The second aspect to a general systems approach to industrial policy that you have to consider is the question of what’s so special about red squirrels anyway. Every politician has their favourite example to hold up. None of them are any good. They might be profitable, glamorous, or anything else for a while, but they are only contingent and temporary. We worry about preserving red squirrels because they are being outcompeted by grey squirrels. But the forest actually doesn’t care. Whilst red squirrels are the most efficient users of feedstock energy, they will have the niche. When something else comes along, it will have the niche. What is key for industrial policy is that it looks at the feedbacks between species as much as it does any one species. Change will happen whether or not it is directed, but a change which cannot co-exist with the system is going to do either of two things: either it won’t take root or it will invade and damage the system around it. How you manage these effects are better strategic choices in general than trying to preserve a system as it is, which may take more energy (=money) to do than can be generated from the preserved system.

A third point is that systems are fuzzy at the boundary. In ecosystems we would be looking at transition zones between systems. For an industrial system, it is recognising that the bounds of the system are open and porous. Entry and exit points have to be kept open enough to keep allowing new sets of entrants to the system or  those who can no-longer exist within it to leave as painlessly as possible. An insular industrial system within an interconnected world doesn’t imply that preservation is worth the effort. It usually implies that you are preventing some flow to another system, making that system behave sub-optimally as well as the system you are trying to preserve.

Finally, the most overlooked feature of an industrial or other system are the ‘hyphens’ within it. Hyphens connect two elements of the system in some way. Intra-system connections are as important as inter-system connectors: they provide essential specialisations which connect, inter-alia, different capital stocks (‘IT in biotech’), expertise (‘technology lawyers’) or people (‘networking’, to the extent that it is actually real does this).

What you emerge with from the systems perspective is an absolute mess of factors that could be control variables if you chose them to be, but would be better off not doing so because their effects on other systems components might be deleterious. Those effects are hidden to you as a ‘system strategist’ working only from a system description. They are often not known to participants within the system (who rely on signals rather than direct outputs). The control variables that make sense are those which improve the quality of the whole system (so good general technological education would be useful. As an aside, I’ve had many reports of what was learnt in IT at schools, and a good many of them were something like “remedial MS office”. How do people learn IT and not C or whatnot? Does the education system really think that the Her Britannic Majesty’s Commonwealth is best served by people who can type a letter when they’re told to rather than people who can hack up a new word-processor if they need to? Don’t answer that. We already know the answer is yes.) and enable transitions from one state of the system to another. (Systems exhibit chaotic transitions. People don’t handle that so well, so gearing unemployment towards managing transitions and enabling people to do the next thing would be useful. That’s just an example for labour. It could be discussed in the same way for changing a quality of capital or a type of input).

Ultimately, unlike biology, economics only takes place in the wild. In the wild, technologies and companies only exist within systems. If we forget this, industrial policy becomes an exercise in futility. If we learn this, we can make fewer and better decisions and be rewarded with more robust and more dynamic systems.