The outlook for capital borrowing in the Scottish Budget

Fraser of Allander Institute, December 2016


One of the central elements of this week’s Scottish Budget will be the outlook for capital spending and an update on the detailed list of infrastructure projects – in transport, housing, schools and hospitals – that the government will take forward over the next few years.

As we set out here, this has been boosted by the Chancellor’s announcement of around £800 million of additional resources for Scotland’s capital budget up to 2020-21 as a result of the measures in last month’s Autumn Statement.

But a major classification decision means that the Scottish Government’s new borrowing powers could – we expect – have effectively been used up this year and a significant amount of the limit for 2017-18 used up as well on dealing with this technical accounting adjustment.

The Scottish Government’s capital spending powers

Under the current fiscal framework, the Scottish Government has the opportunity to fund capital projects – i.e. investment in new schools, new roads and hospitals – through three key sources.

Firstly, there is the traditional capital budget allocation – or CDEL – from Treasury. The Scottish Government’s annual CDEL is determined by the Barnett Formula and is driven by the increase (or decrease) in comparable spending in Whitehall departments. In 2016-17 it is valued at around £3.2 billion.

Secondly, one of the major new powers of the Scotland Act 2012 is the opportunity for the Scottish Government to borrow to fund capital expenditure subject to a statutory aggregate cap of £2.2 billion. An annual limit was imposed of 10 percent of the CDEL allocation in a given year, giving a maximum borrowing limit of approximately £320 million in 2016-17.

Following the Scotland Act 2016, these powers have been extended – with a new statutory limit of £3 billion and annual limit of 15 percent of the overall borrowing cap, equivalent to £450 million.

Thirdly, the Scottish Government has taken forward its Non-Profit Distributing (NPD) model, in part, as a means of topping up capital investment at a time of significant budget restraint. The NPD model was designed to allow for infrastructure to be paid for out of a commitment to make regular payments from revenue budgets rather than an initial one-off capital investment (as is the case under CDEL or borrowing). Essentially a private contractor would take responsibility for the upfront capital cost and delivery of the project (and usually day-to-day maintenance of the assets once the initial project had been delivered) with the government making payments over a number of years (usually decades).

The NPD model was seen as an improvement on the original PFI model used by previous Scottish administrations as it had the advantage of capping the profits private companies could make.

The plan had been for these three elements to work together to support capital investment in Scotland.

ONS reclassification

However, a new system of accounting classifications – the European System of Accounts, ESA10, which took effect on 1 September 2014 – has fundamentally changed the landscape anyway and thrown a particularly large spanner in the works.

In a ruling, in July 2015, the ONS concluded that the Aberdeen Western Periphery Route with a NPD value of £469 million had to be classified as on-balance sheet according to National Accounts classifications.

This may seem all a bit technical and un-interesting. However, it has highly significant implications for the Scottish Budget.

The Scottish Government faces strict rules about how it can fund ‘on-balance’ sheet capital projects – either through its CDEL or new capital borrowing powers.

The effect of being told that the AWPR was now reclassified as ‘on-balance sheet’ meant that the Scottish Government needed to find capital budget cover at the point of initial investment. In other words, they had to turn to their other two sources of capital funding – CDEL and borrowing – instead and could no longer fund it from a commitment to make long-term revenue payments.

In the end – and as we pointed out here – an accounting adjustment was agreed with HM Treasury in 2015-16 so that the capital cover required to support the delivery of the AWPR that year effectively counted against virtually the entire annual borrowing limit for that year (around £283 million).

Recognising the capital value up front, as the classification requires, reduces the future impact on the Scottish Government’s revenue budget.

But by preventing the government from using its borrowing powers to fund other projects, it effectively reduced its capital spending power in 2015-16 compared to what would have been the case prior to the classification decision.

The outlook for 2016-17 and beyond

As set out in the Scottish Government’s Consolidated Accounts for 2015-16, in view of the similarities in the structures of the funding models – a position confirmed by the ONS  – the Scottish Government has adopted a similar budget treatment for three further projects – the Edinburgh Sick Kids Hospital, Dumfries and Galloway Royal Infirmary and the National Blood Centre.

The Hub framework – another revenue finance scheme is unaffected.

To be clear, the Scottish Government has always been up front and clear about this classification change in its various updates to stakeholders.

The projects affected are significant. The NPD component of each project as set out in evidence provided to the Scottish Parliament’s Public Audit Committee last month are –

  • AWPR: £469 million (operational/service start December 2017)
  • Edinburgh Sick Kids Hospital: £150 million (operational/service start April 2018)
  • Dumfries and Galloway Royal Infirmary: £213 million (operational/service start December 2017)
  • National Blood Centre: £33 million (operational/service start March 2017)

This is a combined capital spend of around £860 million to be delivered over the period to April 2018.

The Auditor General’s report of October 2016 suggests that around £328 million of capital budget was freed up from 2014-15 and 2015-16 budgets to accommodate the change in those years (including the use of borrowing powers set out above).

But clearly that leaves a significant balance outstanding.

It follows that for these projects to be delivered during 2016-17 and 2017-18 – which the government is clearly committed to – a similar deal may have to be (or may have been) agreed with HM Treasury as happened in 2015-16 to use the borrowing powers as cover.

A back-of-the-envelope calculation from balances outstanding (around £500 million) means that this would equate to the full amount of the borrowing limit for 2016-17 (around £320 million) and a significant chunk of  the limit for 2017-18 (around £450 million).

Of course, the government may have decided not to use its borrowing powers as additional spending beyond its CDEL limit and NPD anyway but, on balance, it is likely that they would have liked to at some level (or at least had the option).

Conclusions

Whether people agree with the rights or wrongs of the accounting rules it is clearly a risk when designing new innovative funding mechanisms. To be fair, these changes came into effect after the projects had been tendered with construction under way.

NPD programmes have a number of advantages, but they can be more expensive to finance than traditional public infrastructure projects. The government will probably now reflect going forward on the opportunity cost of NPD vs. other forms of funding if both are to be on-balance sheet. This may have happened anyway in the light of the Parliament’s new borrowing powers which have added a lot more headroom than was available when the NPD programme was first devised.

So in this week’s Budget, Mr Mackay is likely to announce an update to the government’s capital plans including arrangements which will effectively take full account of the classification decision. The Autumn Statement’s capital consequentials and extended borrowing powers will help but only in terms of providing support for existing plans rather than anything new.

NB: An article discussing the issue also appeared in the Guardian today – this blog was drafted separately and without prior knowledge of its publication.