24 July 2012
Is it possible to determine the size of a pervasive industry?
Electronics is everywhere: it's a pervasive technology which either enables almost all of the devices we use or makes a substantial contribution to its functionality. But because electronics is pervasive, it's also hard to determine the extent of the industry – particularly in the UK.
There has always been the belief that electronics activity is found in places beyond the most obvious. But how far does it permeate and what is the benefit to the UK economy? That's a challenge being addressed by one of the ESCO workstreams. Leading the activity is Ian Phillips, principal staff engineer with ARM.
He said the research grew out of earlier work by the Electronics Innovation and Growth Team and by the Electronics Leadership Council. "In a way, it was born out of frustrations," he said. "Both bodies had a traditional view of electronics and this didn't really reflect what was going on."
The aim of the ESCO report is to highlight the importance of electronics to the UK, but without a solid grasp of the underlying economics, its impact may well be diluted. So how do you measure the length of the proverbial 'piece of string'? Phillips said: "I thought it couldn't be that difficult; even if it wasn't definitive. As long as it was a repeatable approach and could produce comparisons, it would serve."
Phillips started with what he called a 'simple model'. "If we take the UK's total GDP and divide by the working population, we get a figure per head. The implicit assumption here is the workforce 'makes it happen'."
After that, the process gets harder. "You then have to determine the size and scope of the electronic systems community. I started using a range of databases and, from them, began manipulating by applying SIC filters and so on." A further refinement came from applying the average figures developed to real employment numbers obtained from The Treasury. "We ended up with something like 15,000 enterprises and 200,000 employees contributing 3% of GDP," he noted.
While Phillips thought that was a 'good' figure, he knew it was based on 'too much guesswork'. "I didn't claim it was the best way," said Phillips, "but it generated interest."
Since then, the work has moved from being based on simple division to having its roots in economic theory. "It's based on the income model," Phillips said, "which is nothing new to academics and neither is it rocket science."
Phillips knew from personal experience the initial results were exclusive. "The electronic systems community does all kind of things; we didn't want to get trapped by existing labels when we think we're inclusive." As an example, he cited ARM. "ARM has, until now, been an uncomfortable fit in the electronics industry; it delivers a software product which enables people to develop complex hardware."
The other problem which Phillips faced is that the electronic systems community is 'rather invisible'. "People are developing complex products by offering their skills into a communal 'pond'," he asserted. "They are doing what they are good at for a range of customers. That's a surprise to some people and they are the ones we're trying to address."
Politicians are high on the list of people to be addressed. "They need to know because what they do relates to the incentives that are available to encourage development. We need to show the sector is important and we are being overlooked today because we have failed in the past to show that in a way the public can understand."
Use of 'electronic systems' as a descriptor is not accidental. "We wanted a term that prompted people to ask 'what's that?'. It's a term used in research proposals and it seems to be bringing the industry together." He noted that since NMI has started using the term, it has attracted new members, such as Jaguar Land Rover and McLaren, who don't consider themselves microelectronics companies. "Microsoft is now a member," he added. "Once you have the name right, people are on the same side. If you stick with older descriptions, people become protective."
Even with what he terms a 'rich database', he still faced the question of where the boundaries lie. "Traditional definitions might be good for classification, but they create their own boundaries. The electronics industry has a problem with SICs and geographies," Phillips said. "If you rely on the 10 or so SICs that define the basic constituency, you can get fairly good numbers. But if you're trying to find the genuine engineering activity in electronic systems, the boundaries are harder to pin down, simply because electronic systems are more encompassing."
He encountered one problem when he turned to keyword searching to enhance the SIC results. "With SICs, you don't have to filter too much. But if you use the FAME database and enter 'chips' as a keyword, you include Golden Wonder. So the search needed to be more sophisticated."
Again, he used ARM as an example. "It was one of the more awkward companies to include and wasn't picked up by the primary or secondary SIC search; it only came up in a keyword search." Yet it would be hard to argue against ARM's inclusion in the electronic systems community.
How wide is the net? "Electronics, software, mechatronics," said Phillips. "Then we go into design and manufacturing. That means metrology, robotics, those making power supplies and so on." But how far do you go down a product's life cycle? "Can we draw a line at manufacturing?," he wondered. "No, because systems reproduction is a component. And we have to look at maintenance, which takes a product to end of life."
The community is bounded by a dotted, rather than a solid, line. "It has to be 'fluffy' at the extremes," he noted.
The process also included what Phillips termed 'cross cutting'. "This allows you to include part of a business, rather than all its activity." Included here are businesses which don't identify themselves as associated primarily with electronic systems, but which have a secondary association. "Companies which supply point of sale terminals to banks may think of themselves as being in the financial sector, but they are supplying electronic systems products."
Because you can't 'claim' all the people working for those companies, it becomes harder to apply a GDP model. "You can find out the number of heads working in electronic systems and add that to the contribution," he said.
He also pointed to companies like Tesco and British Airways. "Tesco employs more than 100,000 people, most of whom have nothing to do with electronic systems. Yet Tesco is a leading innovator in logistics – which is electronic systems – and there's maybe 100 people involved. Then there's the IT community, whose design and installation work is supported by work in the electronic systems domain."
Add it all up, he says, and there are 'hundreds of thousands of people'. "Now we have a idea of the value that companies are contributing."
One of the problems, says Phillips, is that previous studies have tended to focus on 'silos'. "That doesn't encourage a feeling of community," he concluded. "There are a lot of companies in the UK doing similar things and we can get value by learning from them. And if they can identify themselves as belonging to a group, it enhances them."