The Fiscal Commission’s forecast evaluation

-

Last week, the Scottish Fiscal Commission (SFC) published its Forecast Evaluation Report.

As with all SFC publications, there was a significant degree of interest in its content.

In particular, it showed that Scottish income tax revenues were some £550m lower than expected. It also addressed the gap between the SFC’s pessimistic forecasts for growth and recent economic indicators.

This blog summarises some of the key issues from the report – with a focus on GDP and income tax. A more in depth discussion (including for the other devolved taxes) will appear in November’s Scotland’s Budget Report 2018.

Background

The documents published last week look at the forecasts made by the SFC back in December 2017 and May 2018.

The forecasts for 2018/19 cannot be evaluated yet as we are only halfway through that financial year.

So the SFC’s Forecast Evaluation Report (FER) looks at the in-year forecasts made for 2017/18 for GDP (and LBTT, Landfill Tax and Non-Domestic Rates).

For income tax, outturn data for 2017/18 will not be available until summer 2019. As a result, the FER evaluates their December 2017 and May 2018 forecasts for income tax revenues in 2016/17.

This transparency is to be welcomed. Albeit, it does leave the Commission open to criticism (some of it justified, some of it less so) if their forecasts turn out to be wrong.

What do they find?

GDP

When the SFC published its May 2018 forecast, it had access to three quarters of GDP data for 2017/18. On that basis it forecast growth for 2017/18 of 0.7% (unchanged from December 2017).

In the end, and according to the latest data published in August, outturn GDP growth for the whole year is estimated to actually been 1.3%. At first glance, this would seem to be a large difference (forecast error).

In fact, the reason for the large discrepancy is largely due to major revisions to the estimates of growth by the Scottish Government.

Indeed, as recently as June, the Scottish Government published growth estimates for 2017/18 of just 0.8%.

Revisions to economic statistics are not unusual, but the scale of these revisions was exceptional.

They were caused by a major change in methodology for construction sector activity – which had been showing some distinctly odd patterns.

The SFC would be correct to argue that, whilst some revisions were anticipated, it is not the SFC’s role to assess the government’s estimates of GDP. Ultimately, the quality of their forecast are constrained to a large extent by the quality of the official data provided. As we have argued before, the challenge is really on the Scottish Government to respond more quickly if the statistics are displaying odd patterns.

At the end of the day, with the revisions largely limited to the timing of activity rather than the underlying health of the economy, the SFC have not changed their expectation of a weak economic outlook over the coming years.

Income tax

In order to understand the differences between the SFC forecast of income tax published in May, and the outturn data for income tax published in July, it is helpful to distinguish two things:

  • The Survey of Personal Incomes (SPI) is a survey of around 2% of income taxpayers and their tax liabilities, published by HMRC. It includes a representative sample for Scotland, based on addresses held by HMRC. The latest year for which the SPI is available is 2015/16. These data form the basis of the SFC’s income tax forecasts released in May.
  • Scottish income tax outturn data for 2016/17 is the administrative outturn data based on the ‘S-codes’ assigned to those with Scottish taxpayer status. This was published in July 2018 for the first time, and comparable outturn data does not exist for any previous year.

When the SFC published its forecasts in December 2017 to inform the 2018/19 budget, there was not yet any published Scottish outturn income tax data for 2016/17. The Commission therefore had to forecast income tax revenues in 2016/17, as a first step towards forecasting revenues in 2017/18 and then 2018/19.

In May 2018, the SFC forecast that Scottish income tax revenues in 2016/17 were around £11.28bn. This forecast was made by taking the estimate of Scottish liabilities in 2015/16 from the SPI, and making assumptions about changes in the drivers of tax growth (e.g. wages, employment etc.)

HMRC has now published Scottish income tax outturn data for 2016/17 showing income tax liabilities were £10.7bn.

This is some £550m below the SFC’s forecast for 2016/17.

The main reason for the large discrepancy does not appear to be because the SFC made wildly inaccurate judgements about the number of taxpayers or income growth.

Instead, it appears to be due to major underlying differences between the SPI survey data and the administrative outturn data. Indeed, not only is the 2016/17 outturn data below the SFC’s forecast, it is also lower than the estimate of liabilities based on the 2014/15 and 2015/16 SPIs.

The SFC explore some potential reasons for this – these include that there may be differences in taxpayers defined as being ‘Scottish’ in the two sources, or that the SPI sample is not as representative as it could be.

In particular, the SFC believe that most of the error stems from the number of higher and additional rate taxpayers being lower than previously thought.

Previous forecasts for income tax receipts assumed there to be 18,000 additional rate taxpayers and 337,000 higher rate taxpayers. The latest data shows that there were in fact 13,000 additional rate and 294,000 higher rate taxpayers.

The reduced number of higher and additional rate taxpayers is estimated by the SFC to account for £500m of the £550m forecast error.

We will need to wait to see how future SPI releases compare to outturn data, starting with the 2016/17 SPI published in Spring next year.

Budgetary implications of the revision

Despite Scottish outturn revenues in 2016/17 being lower than forecast, this in itself is unlikely to affect the resources available to the Scottish budget in 2017/18 or any future year.

This is because 2016/17 is used as the baseline for adjusting the Scottish block grant for the new income tax powers. The lower revenues are, the smaller the ‘initial deduction’ is. What matters for the Scottish budget in 2017/18 is how quickly income tax revenues grow after 2016/17 in Scotland compared to the rest of the UK

A lower than forecast outturn figure does not tell us anything about the relative growth rate of Scottish revenues in the future. Yes, it will mean that the SFC revises down its forecast of Scottish revenues for 2017/18 and beyond. But the block grant adjustments for income tax will be revised down too.

However, there are some important implications to consider.

In particular, if the outturn data is correct, it would appear that Scotland has fewer higher and additional rate taxpayers compared to rUK than was previously thought.

What might be the implications of this?

On the one hand it could mean that it is less likely that the growth of Scottish income tax revenues per capita will match the growth of rUK tax revenues per capita.

Why?

  • Consider what might happen if a larger share of Scottish taxable income is earned at the basic rate of income tax, compared to rUK. In this case, UK-wide factors which reduce taxable income at the basic rate have more of an impact on total tax revenue in Scotland compared with rUK. For example, a real terms increase in the Personal Allowance – as has been happening recently – reduces Scottish revenues proportionately more than it would reduce rUK revenues (and hence the BGA).
  • Similarly, if a smaller share of Scottish taxable income is earned at the additional rate, UK-wide factors which disproportionately increase taxable income at the additional rate would have more of an impact on rUK tax revenues than Scottish revenues. For example, if the incomes of additional rate taxpayers grew faster than the incomes of basic rate taxpayers, in both Scotland and rUK, this would have a proportionately greater impact on total revenues in rUK, given that income taxed at the additional rate makes up a larger share of rUK tax revenue.

Of course these differences in the distribution of taxpayer incomes could work in Scotland’s favour in some circumstances. These distributional issues demonstrate why, under the Welsh Fiscal Framework, block grant adjustments for income tax are calculated separately for each of the three bands of rUK income tax.

In summary, it was always assumed that the SPI would provide a relatively robust estimate of Scottish income tax revenues in previous years, and thus form a good basis for making forecasts. It now appears that the SPI aligns less well with outturn income tax data than had been thought. How this will play out in the future remains to be seen.

Summary

The SFC concludes that its forecasts are ‘reasonable’.

Whilst ‘reasonableness’ is a subjective term, the SFC’s in-year forecasts don’t appear to be too bad, once allowance is made for data issues.

But the consequences of SFC forecast error will become much more tangible in 2018/19 and beyond, as it is the SFC’s December 2017 forecasts for 2018/19 that determined the size of the Scottish budget in that year. There will be much greater interest in the scale of these ‘year ahead’ errors.

What this has highlighted is some of the highly complex technical issues that were always going to plague the transfer of major new tax powers to the Scottish Parliament with only limited data and intelligence on the Scottish economy or taxpayers.  And this is before the much more complex issues of VAT and social security payments have been factored in!

September 10, 2018