A bespoke statistics agency could show us what's really happening in Scotland

More seriously, this problem plagues just about anyone who tries to rely on national statistics for research or policy-making purposes. This is as true for the government as it is for the think-tanks and lobbyists who try to guide or influence them. No-one wants to find out that a keystone government policy is based on bad or absent data and few would be happy to find out that the impact that a policy is having on society is being tracked inadequately or not at all.

READ MORE: Could this be the solution to the annual GERS controversy?

I’m not going to criticise directly the work of the civil servants behind the gathering of statistics for the UK and for Scotland. Contrary to the apparent belief of a certain hyphenated MP last week, they overwhelmingly gather and publish data impartially and leave it to others to apply their desired spin on the numbers. However, it is clear that critical gaps exist and work could be done to fill them.

Common Weal’s latest policy paper began as an exploration of how an independent Scotland could build and manage its own independent Scottish Statistics Agency (SSA) but over the course of a year of researching, analysing and talking to people in the sector at both Scottish and UK levels it became clear that not only could the Scottish Government do a lot of this now, before independence, but that there are good reasons for it to do so.

It can be said that the provision of data in Scotland is better than in the UK as a whole but one response to the Scottish Government’s recent consultation on data described our statisticians as “opportunistic, but relatively powerless, statistical scavengers”.

At a UK level, statistics are often gathered from the point of view of looking at the UK as a whole. This means that if someone wants to look at what is going on in Scotland, they have to start “sub-sampling” the data. Let’s say that the UK does a poll of income distribution across the UK and it asks 10,000 people how much they earn.

This is a reasonable sized sample for a poll but if we want to use it to look at income distribution in Scotland then we may find that only 830 people have been asked up here. This might still be fine for a broad overview but that’s only 26 people per local authority and we haven’t even tried to break the numbers down by gender, age or other factors yet.

Scottish statisticians sometimes try to get round this by “oversampling” Scotland to bring the numbers up but this clearly just pushes the limit further out to the edges rather than dealing with the core problem itself. Sampling is always limited and we could be looking at different ways of gathering data instead, such as via tax records. To take another example, the UK’s data on trade is fairly thin. In Scotland we have some data on Scotland’s goods exports – hence the annual headline that “the UK is Scotland’s largest trading partner” when the Export Statistics Scotland (ESS) report publishes.

The thing is, this report is based on a very small survey of Scottish companies and ESS can’t even compel them to fill out the form (response rates are less than 30 per cent). Whilst it can, in most cases, rule out the “English Ports” problem (goods going to France are counted here as a Scotland to France export even if they leave the UK via a port in England), we’re lacking something rather more fundamental. We have little idea what Scotland (or the UK) imports. The UK has gutted its customs capability in favour of focusing on immigration and shows no sign of changing this policy even as it tries desperately to hurtle itself out of the EU Customs Union. On services, it’s even worse. A little reported story last week revealed that if you ask the UK what its services trade balance with the rest of the world is, it’ll say it has a $77 billion surplus. But if you ask the rest of the world for its figures they say that the UK has a $39 billion services trade deficit. Someone is wrong. We don’t yet know who.

And don’t get me started on GERS...

I’LL encourage readers to dive into the full report for more tales of data gaps and limits as well as our recommendations of what can be done in each case. We think that a lot could be done with a bespoke SSA which could found itself on principles not just of impeccable data gathering but also dissemination as well. It should be possible for you to go to a single data portal on the web and basically set your parameters for the data you need.

The SSA’s job would be to ensure that the data is gathered and processed in such a way that it can be easily and transparently delivered and different datasets can be combined easily to avoid errors creeping in. We could also set up a system of “kitemarks” which state that a particular set of data has been gathered to a standard high enough for use in policy-making decisions. This is similar to the UK’s Code of Practice system but could be expanded to allow it to be awarded to non-governmental data bodies like academic researchers and think-tanks – if they can meet the standard.

There are, of course, multiple ways to build an SSA – it could be centralised into one organisation or decentralised across specialised departments – but one thing it should be is substantially larger than current provision. An agency the same proportional size as the UK spends on statistics would employ 265 people in Scotland. Sweden employs 700 people in their statistics agency whilst New Zealand employs over 900. These are highly skilled jobs and the comparatively modest costs of the agency would pay for itself in the form of more effective, better targeted, more ably tracked policy. And who knows. It might even stop some of the arguments on social media too.