Bill Jamieson: glaring errors by the Scottish Fiscal Commission undermine confidence

A figure hailed as a crack shot enters the shooting gallery. It's the Scottish Fiscal Commission. The rifle is cocked and an expectant hush descends. There is a spray of bullets. Unfortunately, none hit the target and instead are splayed over the ceiling and walls.

What to do? Get a new rifle? Reload and leave the building? Or send the SFC to SpecSavers?

There really is no easy solution to the problems laid bare by the latest SFC report – the body set up to provide a reliable, independent check on the Scottish Government’s tax revenue estimates and its figures on economic performance.

Instead there is a litany of grievously inaccurate forecasts, estimates missed by mile, wordy excuses and unconvincing promises to do better in future. On this evidence, future reports may need the helping tap of a blind man’s stick.

The £50 million overestimate of revenue raised by the controversial business rates levy is the least of it. Nor need we worry much about the £13m overestimate of revenue raised by the residential Land & Buildings Transaction Tax, due to over-optimistic projections on transaction numbers.

No set of projections can be 100 per cent accurate. And these errors fall within a reasonable margin of error. But when we come to the major estimates of tax revenue and overall economic performance, the errors are glaring and merit scrutiny. For it is forecasting in these areas that backlight key decisions by the government of the day on tax and economic policy.

On economic growth, the SFC forecast as recently as May that Gross Domestic Product – our main indicator of economic performance – would grow by 0.7 per cent in 2017-18. But the latest data release shows growth of 1.3 per cent, or almost double the SFC estimate.

This is a worrying degree of error by any standards, and particularly so for a body set up to blow the whistle on inaccurate numbers from other sources. The SFC waives this away, saying that “our GDP forecast error is within what we believe is a reasonable range based on the track record of the Office for Budget Responsibility (OBR) forecasting the UK economy” – a low bar given the OBR’s own errant record.

The SFC’s explanation is that the inputs it relied upon were faulty – in particular, “significant revisions” to historic GDP growth, due to a major rewrite of construction industry activity since 2015. GDP growth in 2017-18 was revised up by 0.5 percentage points – from 0.8 per cent to 1.3 per cent. In the words of the SFC, this was “an exceptionally large revision by historic standards” – around five times what is typically the size of revision to annual GDP growth between publications.

Latest figures now show that the construction industry, rather than falling by 12 per cent since 2015 as originally thought, grew by 4 per cent – a dramatic rewrite of data which was the main reason for the SFC’s forecasting error of 0.55 percentage points for 2017-18.

You see what I mean by the demonic splaying of bullets: hefty warnings had been made at the time by independent economists John McLaren and Tony Mackay when the Scottish Government first regurgitated the construction figures, but little notice was taken at the time. The worrying issue here is that if the SFC is so dependent on Government faulty data, what hope can there be that future estimates and assessments will be any better?

However, the most glaring forecast error is to be found in the SFC’s projections on Scottish income tax revenues. In May it estimated these would total £11.3 billion. The HMRC outturn data now puts the figure at £10.7bn. The SFC explains that its estimates were based on an “anonymised and publicly available” version of an administrative data sample held by HMRC. This overestimated income tax liabilities and is the main reason for the SFC’s income tax forecast error of £550m – “a greater forecast error than the benchmark OBR average UK income tax forecasting error we have calculated”.

Any the wiser? It appears the discrepancy may be due to differences in the estimated number of taxpayers with higher levels of income, and in particular differences in the number of additional rate taxpayers. Indeed, this could account for some £500m of the shortfall.

So, wonky data, yet again. But for a Scottish Government feeling its way on changes to starting rates and levels for higher and additional rates of income tax, more accurate numbers are vital. Just how many taxpayers are liable for higher rates? Claims in the parliament chamber or on the hustings that some new spending measure can be easily afforded by adjustments to taxes levied on the “better off” ring hollow in the face of a discrepancy that can extend to £550m – greater than the budget of the country’s enterprise agency.

This is especially alarming given the warnings at the time – in this column and elsewhere – of the likelihood of behavioural responses to changes in income tax rates – higher and additional rate taxpayers opting for example to up their pension plan contributions.

Still to come are estimates of the apportionment of VAT revenues to be assigned to Scotland – a statistical terrain about as capable of accurate estimation as the shifting sands of the Sahara. Retail revenues, tourist estimations, small business exemptions, digital-based services, internet-based spending: problems here could dwarf those incurred on income tax estimates.

Meanwhile the Scottish Retail Consortium lost no time in pointing out that if Holyrood wants to see higher revenue from Scotland’s share of VAT, improving levels of household income and spending are vital.

I do not envy the SFC one jot in seeking to improve the quality of the input data on which it relies. Even making generous allowance for margins of error that accompany any economic forecast or revenue projection, major improvement here is urgently needed.

“The further ahead we look,” runs a particularly platitudinous sentence in its summary, “the more uncertain the future”. Well, I never. Those setting out on the journey through these windy 83 pages need not struggle to the end to arrive at conclusions over uncertainty in the “further ahead”. It’s the forecasting of what lies immediately ahead that’s troubling enough.