In the UK regions ‘Whitehall’ has long been a by-word for scepticism about policy created in faraway London. It can be loaded with the idea civil servants in the capital have a limited grasp of the economic development challenges and social issues outside of the M25.
Whitehall also has a symbolic place in British history: it was outside the Banqueting House on Whitehall that King Charles I met his end. On a cold January morning in 1649 the assembled crowd is reported to have groaned as the axe came down. Although the monarchy was later restored, the execution destroyed the idea of an all-powerful and unquestionable monarch.
The view of Whitehall as a bastion of the UK establishment endures to this day, despite recent strides to devolve power to Metro Mayors. Its reputation is closely related to the quality of policy-making happening at the centre of government and here, in my view, there’s obvious scope for a new renaissance.
The main ask is the application of 21st-century data science. Accurate, real-world data needs to underpin any and all decision making and get us away from a world where there are numerous examples of wonky data in action. Everyone talks about the exploitation of Big Data. If we are to have a meaningful, data-led approach to policymaking in the UK (using big or small data), it can only be successful if that data is accurate.
One of the most pressing concerns relates to the Standard Industrial Classification (SIC), a system for classifying industries by a four-digit code, which is in desperate need of an overhaul.
Established in the United States in 1937, SIC codes are used by UK government agencies and Companies House to classify industry areas. The system was designed to overcome the problem of branches of government conducting analysis using a variety of methods and metrics, unknown and meaningless to other departments. SIC codes were to bring order and uniformity, with businesses grouped around common characteristics shared in the products, services, and production.
They have a hierarchical, top-down structure that begins with general characteristics and narrows down to specifics. The first two digits of the code represent the major industry sector to which a business belongs. The third and fourth digits describe the sub-classification of the business group and specialisation, respectively.
It sounds perfectly logical but in practice the integrity of the system has been long since been lost to the advance and increasing complexity of commerce. The system has always been vulnerable to the fact that companies self-select what category they belong to, and are responsible for maintaining the accuracy of their classification. In reality, once a SIC code has been selected, it is very rarely changed. It means that, some 80 years after their introduction, we have a free for all.
To take one example, the category that pools companies in scientific research and development is full of absurdities. Springwell Microelectronics Ltd, a manufacturer of sensors which activate toilet flushes, is listed amongst them. Similarly, Goat Nutrition Ltd, which supplies small holders and goat keepers with equipment is identified under ‘Manufacture of Pharmaceutical Preparations.’ Better still, Glaxosmithkline plc, one of the global giants of pharma, has a listing that makes no reference to development of new drugs, but is divided between ‘Manufacture and Wholesale of pharmaceuticals’, ‘Other Professional, Scientific and Technical Services’ and Other Business Support Services. Their research subsidiary classes itself as ‘Other Research and Development on Natural Sciences and Engineering.’
Another problem is that SIC codes put sectors together in odd ways. Biotech and software development are based on intellectual property. But biotech belongs to a sector group called ‘Professional, scientific and technical activities’, which means it is in the same group as Public Relations. And software development belongs to a sector group called ‘Information and communication, which includes sectors like ‘Motion picture distribution activities.’
These issues are manifold across the 600 and more SIC codes. This matters because Companies House data informs understanding and policymaking in Whitehall and beyond.
Over the last two years, LEPs and Combined Authorities have been engaged in developing their Local Industrial Strategies with the specific intention to promote the coordination of local economic policy and national funding streams and establish new ways of working between national and local government, and the public and private sectors. Most of these groups have sought to use data on their business base to support their developing strategies and justify funding. This approach is confounded by inaccurate and unreliable data that will simply result in the inaccurate allocation of resources to strategies whose underlying assumptions are fundamentally flawed.
Our economy has evolved from one that was based on manufacturing, to information and service-based sectors. SIC codes have not adapted quickly enough with the last major revision taking place in 2007. Instead, vague catch-all’ classifications have been introduced such as ‘Other service activities’.
As part of that approach, there needs to be a fundamental reform of how SIC codes and used and policed. It should go without saying that SIC codes were designed to bring order and uniformity to aid in understanding what was going on in our business base – we have moved a long way from that original ambition.
We need better education about how to use SIC codes when a company is formed or when its core business activities change. In an ideal world, a reformed system would also include some reference to business value in the classification system. Most critically, the new system should be policed to ensure that it remains current with penalties for companies that do not comply. That way, perhaps we can get back to a system that will actually help, rather than confound policy making.
Impact Data Metrics