In late May 2014, the UK Treasury released a press release ahead of a major report UK minsters were issuing on the costs of setting up an independent Scottish state. Joel Suss, Managing Editor of British Politics and Policy blog, asks Patrick Dunleavy about the way in which the Treasury used his research findings to arrive at a figure of £2.7 billion, and about his Twitter intervention raising concerns about it, which caused a furore. How did this episode come about, and does it serve to illustrate the problems of mis-communication between civil servants and researchers?
What were the main problems with the press release from the Treasury on the reorganisation costs of an independent Scottish Government?
The UK government correctly said that in our 2010 LSE Public Policy Group report Making and Breaking Whitehall Departments, Anne White and I estimated that the cost of setting up a new medium-sized department at UK national level was then £15 million. This covers just the extra expenses of reorganisation itself, and not any possible enhanced effectiveness gains. It is (if you like) the transaction cost of change (like paying a solicitor to do your house conveyancing).
The Treasury then took this £15 million estimate and said that an independent Scotland government would need to create 180 new bodies. (I’m not clear how they got that number.) Next they multiplied £15 million by 180 to get £2.7 billion as the ‘set-up’ costs of an independent Scotland’s government.
This is a very crude series of steps to make. Alex Salmond has rightly pointed out that the Scottish government are not proposing to create 180 Whitehall-scale departments in Scotland. Whitehall departments are very big and expensive organizations, the Rolls Royce of administrative machines. The Secretaries of States, Ministers of State, junior Ministers and Permanent Secretaries all have large staffs and costly offices. Highly siloed Whitehall IT systems need to be altered and reconfigured in any reorganization. Offices need to be moved. Often staff contracts and pay levels from different departments have to be unified. And everything needs to be rebranded. These costs all add up. So our estimate was based on the size and function of an average Whitehall department reorganization.
Actually, many of the new bodies needed in Scotland could be very small. For example the UK Electoral Commission has an annual budget of £21 million for its work (covering a 61 million population). You certainly would not need to spend £15 million to create a Scottish Electoral Commission, covering only 5.5 million people.
How could the Scottish Government go about creating a new government structure and what do you think that would cost?
If you look at the Scottish Government website you can see that it doesn’t now have a highly siloed department structure like Whitehall. It currently uses the same kind of structure as the European Commission, where you have a set of more flexible ‘directorates’ with common IT and HR systems. This is a more modern and cost efficient way of running things than Whitehall’s legacy departments. So I could see that the Scottish government might want to retain that way of doing things for much of its work after independence. Even if it didn’t do so, across the OECD countries the average number of ministries is 14. So I certainly don’t see why Scotland would need any more top ministries than that.
Over and above the structures already in place in Edinburgh, an independent Scotland would clearly need four extra ministries. Two would be brand new: a ministry of defence, and a ministry of foreign affairs (together with embassies, which could be a bit expensive). Two more would involve reorganizing and adding to the work already done in Scotland on tax collection (by HM Revenues and Customs (HMRC)), which might form part of a new Ministry of Finance; and a welfare ministry, taking over what Department of Work and Pensions (DWP) already has in place in Scotland. When the Financial Times asked me to estimate the reorganization costs for enlarging the Scottish Executive directorates, plus these four additional ministries, that is where I guesstimated a number around £150-200 million.
The Treasury press release also estimated that an independent Scotland would face extra IT and administrative costs for tax collection of ‘as much as £562 million’ and for ‘a new benefit system … £400 million alone’. Were these included in your estimate? And do they look right to you?
Our original 2010 research only covered Whitehall reorganizations, not whole scale new policy systems, so these elements were not included in my guesstimate above. Again, I can’t really comment on the provenance of these large Treasury numbers, and so I don’t know how seriously to take them. Both the Department for Work and Pensions and HM Revenue and Customs already have offices in and systems for Scotland, so you wouldn’t need whole new organisations on the ground.
Where costs could be substantial is in disentangling Scotland’s tax and welfare IT systems from those of the rest of the UK. Currently both HMRC and DWP have very complex ‘legacy’ IT systems, with a great many mainframe computers systems arranged in complex networks, and a lot of old-fashioned software code that is expensive to maintain and operate. It is conceivable that an independent Scotland would have to undertake a lot of work quickly to pull out the data that relate to Scotland only, and to get new IT systems in place. (By the way, the rest of the UK would essentially have to do the same work, e.g. identifying which people were Scottish and which rUK citizens. So some set up costs might be shared.)
However, if a new Scottish government was to spend sums as great as the Treasury brief suggests, it is also inconceivable that it would do so without trying to significantly improve the tax and welfare IT systems, so as to make both productivity gains and efficiency savings, and to create more effective systems. And in fact creating a new IT system for a small country like Scotland could be made much cheaper than current UK large legacy IT systems that are costly to run and maintain. For instance, I would estimate that the current Swedish tax system is about 15 years ahead of the UK in its use of digital tax collection. So although it might cost to create new tax and welfare systems at the start, I guess that if you were investing substantially in new IT you’d be looking at systems yielding long run cost savings over 10 or 20 years. Over time running cost savings could significantly off-set some of the up-front costs of a change. (To help think about this, let’s take an example. What would I do if someone suddenly took away my old Dell desktop and I had to replace it? When I go down to the shop, will I buy an exact clone of what I have already? Or will I use the spending opportunity to invest in a new much better PC?)
Are there other estimates for reorganising or creating new government bodies that might have been more suitable?
Well, there was a National Audit Office report published at the same time as ours (March 2010) that looked at the costs of UK level reorganisations below the Whitehall department level. Their analysis fits more closely to what Scotland would have to do on agencies and public bodies in the event of a ‘Yes’ vote.
How would you overall evaluate the information that the Treasury has issued?
I think that UK ministers and the Treasury potentially have a point in arguing that the Scottish Government has not been very forthcoming, they might even say evasive, about clarifying some of the costs entailed in independence. But there is no point in putting into public circulation misinformation to try to counteract that. My objection is that the Treasury woefully misapplied our research estimates and that this did not need to have happened. If officials had rung me at any time (as e.g. any good journalist would) I’d have been very happy to brief them on our findings. and on the issues involved in producing good quality information here.
I’ve actually just published a book (with co-authors) on how academic work has an impact on policy-making. So this is an interesting little incident to add to the picture, illustrating the problems of mis-communication between civil servants and researchers. It shows how people in government need to be a bit more engaged with academia in an effective (and not exploiting) way. It also shows how policymakers need to consider now the power of social media, such as Twitter. I simply sent two tweets the afternoon I spotted the Treasury mistake [one of them shown below], which were then picked up, and today the story made the front page of the Financial Times.
UK Treasury press release on #Scotland costs of government https://t.co/5prEDjJJsZ badly misrepresents LSE research http://t.co/41QX5H4bwa
— Patrick Dunleavy (@PJDunleavy) May 27, 2014
Note: This article gives the views of the interviewee, and not the position of the British Politics and Policy blog, nor of the London School of Economics. Please read our comments policy before posting. Homepage image credit: Tom Page
About the Interviewee
Patrick Dunleavy is Co-Director of Democratic Audit, Chair of the LSE Public Policy Group, and a Professor of Political Science at the LSE.
OK, let’s take the £1.5bn estimate at face value
Scotland currently pays £1.66bn a year into the UK’s coffers, or rather towards the Uk’s debt.
I can live with £1.5bn set up costs, and Professor Dunleavy’s lower estimate would be recouped in weeks.
Hmmm – to be fair there is a distinct level of selectivity here
The Treasury paper quite carefully set out the fact that the figure was not their own prediction of cost, but published a range of figures from different sources – ONE of those figures, the highest, was the one from Dunleavy. The treasury quite carefully pointed out that this was nothing more than a simple extrapolation of the Dunleavy average figure and the Scots governments own estimated number of departments involved, and not the treasuries own worked prediction of cost (in fact they indicated that 1.5 Billion was a more realistic figure).
I also find prof. Dunleavys defence somewhat confused – “You certainly would not need to spend £15 million to create a Scottish Electoral Commission, covering only 5.5 million people.” – Well, no, of course not, but your paper calculated an ‘average’ figure, some departments will be bigger, some will be smaller, pointing to one example of a smaller figure is specious, as its bound to be the case that some cost less, and some cost more, its the very nature of averages surely?
For clarity – the treasury text is below, I fail to see how anyone reading it could see it as anything other than a simple and non-authoritative working, trying to pretend it was some complex misrepresentation or misapplication of the figures is silly, it was just one of a range.
“The Institute for Government (IfG) and the London School of Economics and Political Science (LSE) estimate the average cost for a new policy department or a mid-sized merger to be approximately £15 million. In Scotland’s Future, the Scottish Government estimate that 300 institutions currently serve Scotland, of which 180 would need to be recreated or have powers transferred to a Scottish institution following
independence. If this cost were incurred for all 180 organisations, the total cost would be
£2.7 billion. Given these estimates, £1.5 billion is likely to be a favourable estimate of the total
costs of setting up new institutions.”
Oh come on! “Scottish Executive”?? The Prof used “Scottish Government” the name that’s above the door. Even Westminster dropped the Executive title after the SNP renamed the organisation. Now it’s only used pejoratively or ignorantly. What is it Joel?
Accidentally in this case! It has now been amended.