Costing the proposed Scottish Income Tax band between the higher and top rates

Today we are publishing a short briefing on how much revenue would be raised by some of the proposals for an additional band of Scottish Income Tax between the higher (42p in the pound) and top (47p in the pound) rates. There are two separate proposals in the public domain:

  • One from the Scottish Trades Union Congress (STUC), which would levy a 44p rate on earnings between £75,000 and £125,140; and
  • One from IPPR Scotland, which would levy a 45p rate on earnings between £58,285 and £125,140.

Estimates from the respective organisations are that these would raise £200m and £257m a year, respectively, but these are static estimates.

Our briefing goes into detail about how we have modelled behaviour – but our approach is very similar to that used by the Scottish Fiscal Commission when it comes to including policy costings in their forecast.

We quantify two main behaviours:

  • The ‘intensive margin’ effect, which quantifies how much declared income changes by at the marginal rate that people face, on the assumption that they continue to have the same income streams. This is called the marginal effective tax rate (METR) effect in the literature, and is generally the largest behavioural effect.
  • And the ‘extensive margin’ effect, that is, to what extent people chose whether or not to participate altogether in an activity given their new average liabilities. For example, people may choose to not work anymore, or may no longer find it worthwhile to rent out a property. This is called the average effective tax rate (AETR) effect in the literature, and tends to be significantly smaller than the METR effect.

Our best estimate is that the STUC proposal would raise £56m in 2024-25 after accounting for behaviour, and would be paid by around 141,000 people (5% of those with any Scottish Income Tax liability). We assume that the £75,000 threshold remains fixed in cash terms, which would bring more people into scope of the new band. We think that would be as many as 189,000, or 6% of those with Scottish Income Tax liabilities. If the threshold were to be indexed going forward, revenues raised would be lower in later years.

In the case of the IPPR Scotland proposal, revenues raised would be larger – both because the band starts at a lower level and because the new rate would be higher. We estimate that it would raise around £161m in 2024-25, and be paid by 234,000 people – roughly 10% of those with Scottish Income Tax liabilities. Under the same assumption of a fixed threshold in cash terms, this would rise to £220m by 2028-29 and the rate would be paid by 372,000 taxpayers – 12% of those with any Scottish Income Tax liabilities.

In both cases, the importance of modelling behaviour is clearly illustrated. The static costing is much larger than accounting for taxpayers’ responses – but these responses are well documented, and therefore need to be considered when studying the merits of a policy proposal. Using the well-established SFC methodology – which is also very similar to that used by HMRC at UK level – behaviour reduces the revenue raised through these measures by an around a third: 30% in the IPPR Scotland proposal and 36% in the STUC proposal. Most of this is due to the intensive margin effect, which on its own reduces the yield of the measures by around a quarter.

While these are still significant sums of money, they are substantially lower than they would appear by assuming no taxpayer behaviour. It is important that their anticipated consequences are fully considered as they are debated in the public sphere in the run-up to the decisions for the Scottish Budget.


João is Deputy Director and Senior Knowledge Exchange Fellow at the Fraser of Allander Institute. Previously, he was a Senior Fiscal Analyst at the Office for Budget Responsibility, where he led on analysis of long-term sustainability of the UK's public finances and on the effect of economic developments and fiscal policy on the UK's medium-term outlook.