Chapter 7 - Non-Domestic Rates Income (NDRI)
Non-domestic rate income forecasts have been produced by the Scottish Government for each draft budget since devolution in 1999. There are many factors which affect NDRi with the key components relating to changes in the underlying tax base, tax rates and reliefs and any subsequent backdating that might occur.
The underlying tax base is made up of every non-domestic property in Scotland that is valued by the Scottish Assessors and documented on the publically available valuation rolls (from which some type of property are exempt). All properties are valued to a common 'tone date' (for 2016-17 the tone date was 1 April 2008, relating to the 2010 revaluation). Revaluations take place periodically (generally every five years under current legislation) and a new tone date of 1 April 2015 will apply to the valuation rolls from 1 April 2017.
Changes to the tax base occur each year due to new properties being entered onto the Valuation Roll, changes to existing properties ( e.g. physical changes or changes of use, which can involve removal from the Valuation Roll) and successful valuation appeals. We classify appeals into two categories: running roll appeals typically occur due to a material change of circumstance that affects a property and revaluation appeals refer to disagreement on the initial value set at revaluation - revaluation appeals are accounted for outwith the buoyancy forecast. Growth in the tax base once revaluation appeals have been accounted for is known as buoyancy.
The amount of NDR that each property must pay depends on the poundage rate and any applicable supplements and/or reliefs. NDRi in a given year is also dependent on collection rates, backdating of appeals and reliefs and any other adjustments that affect income such as Tax Incremental Finance projects.
For Draft Budget 2017-18, the SFC has responsibility for scrutinising the economic determinants of the forecast, which extend to inflation and buoyancy.
Inflation forecasts typically feed into the uprating of the poundage rate each year although the decision on the level of poundage is for Scottish Ministers, subject to parliamentary approval. In recent years, the Scottish Government has adopted the same poundage rate as the UK Government. The UK Government typically uprates poundage each year by September RPI of the preceding financial year ( e.g. September 2016 RPI is used to uprate 2017-18 poundage). From 2020-21 the UK Government have indicated a change of policy to use September CPI rather than RPI. Our inflation forecasts mirror this adjustment, and we use the OBR figures published at Autumn Statement 2016 as the best available evidence - an approach which the Commission have confirmed as reasonable.
This extends to inflation forecasts where they are used for the purpose of the forecast, and expected growth in the rateable values ( RV) of the underlying tax base once revaluation appeals have been taken into account - this growth in RV is otherwise referred to as "buoyancy". From April 2017, the Commission will also be responsible for an overall NDRi forecast. However, the forecast methodology included in this paper only covers the parts of the forecast under the Commission's current remit.
Forecasts of buoyancy
The Scottish Government forecasts for buoyancy are as follows:
Table 28: Buoyancy Forecast (%) as at Draft Budget 2017/18
|Assumed Year of 2017 Revaluation Cycle||1||2||3||4||5|
|Long Term Average Buoyancy||1.2||1.2||1.2||1.2||1.2|
|Buoyancy Forecast ||1.7||1.8||1.1||0.9||1.0|
Only the 2017/18 forecast is used as part of Draft Budget 2017/18. The remaining figures are provided only to illustrate the methodology used.
The rest of this chapter focuses on the modelling approach for forecasting buoyancy including the limitations and some basic sensitivity tests to understand the risk of deviation from forecast.
The key challenge with buoyancy forecasting continues to be the lack of data, with just 15 reliable outturn data points on buoyancy available over a 20 year period. Buoyancy is a relatively stable series and we have not observed large fluctuations in growth. However, we have noted minor fluctuations around the long run average. For this reason, our approach to date has been to use the long run average of the outturn series with adjustments made if we have sufficient evidence.
In the run-up to Draft Budget 2016-17, following challenge from the Commission, the Scottish Government started to analyse whether a cyclical pattern could be partly responsible for the small fluctuations in buoyancy during the revaluation cycle. The cycle is apparent in visual inspections of the data with higher buoyancy in the early years of a cycle then in the later years on average. This is shown in Chart 17 with dotted vertical lines denoting revaluation years.
Chart 17: Buoyancy outturn as known at Draft Budget 2016-17 (data not available for 1995-96, 1999-00, 2000-01, 2004-05, 2005-06)
We have been unable to confirm the existence of a cyclical pattern using statistical tests  but through discussions with the Scottish Assessors, we determined that delays and subsequent backdating of running roll appeals could be causing a cyclical pattern in buoyancy. In the run-up to Draft Budget 2016-17, some data analysis was taken forward that provided quantitative evidence to support this explanation. However, the scope of this analysis was limited.
To better evidence the observed cyclical adjustment for Draft Budget 2017-18, we filtered the Valuation Roll to identify each of the 100,000+ changes made to any entry over the period 2010-11 to 2015-16. By comparing "effective dates" (the date at which a change was made to a property) and "valuation notice dates" (the date that these changes were recorded on the Valuation Roll) we were able to identify a cyclical effect consistent with the previously observed cyclical pattern. The data supported the theory that changes in RV which were delayed as a result of the lead-time between running roll appeals being raised (which references the effective date) and running roll appeals being resolved (the valuation notice date). However, because running roll appeals cannot be systematically identified within the data, there is not definitive evidence that running roll appeals are causing this cyclical effect - even if it would appear to be a reasonable explanation.
We further cross-checked the results of this analysis with a subset of running-roll appeal records from the Renfrewshire Valuation Joint Board  . This work confirmed that delays to running-roll appeals here were the same as our effective date/valuation notice date analysis had indicated.
As a result of these two pieces of analysis, the evidence base supporting the use of a cyclical adjustment has been strengthened considerably over the course of the year - and we have chosen to continue to apply the same cyclical adjustment that was applied in Draft Budget 2016-17.
New data for 2015-16 also gave us information regarding the suitability of using the long-run average as the buoyancy figure for this year was higher than the current forecast methodology would have implied. As noted in Chapter 2 on outturn, we were able to identify the reason for this higher than average figure as being down to two particularly large entries to the Valuation Roll (+£10m RV)  .
As a result of this, and following on from the recommendation from the Commission of the importance of tracking large projects in development, we sought advice from the Scottish Assessors Association who routinely monitor planning information and progress of developments to ensure that new properties are added to the Valuation Roll at the appropriate time. They have not indicated that they know of any RV projects in excess of £10 million in the near term. As a result, we have continued to use the long run average as part of our forecast methodology.
We still lack enough data points to be able to run robust statistical tests both in relation to the cyclical adjustment and on any relationship of the cyclically adjusted outturn data to economic conditions. It will be a number of years before this is possible.
Whilst we judge evidence we have used to explain the cyclical adjustment to be extensive we weren't able to systematically identify running roll appeals as opposed to general changes to the valuation roll. Therefore we have assumed that changes to the valuation roll with effective dates in a different financial year from valuation notice dates to be running roll appeals. Our efforts to verify this with data for Renfrewshire was helpful, but we could not extend this to the whole of Scotland.
In addition, whilst we have observed a similar cyclical pattern over four revaluation cycles, we cannot conclusively predict that the same cycle will repeat in future.
To try and produce a meaningful guide to the impact of potential forecast error we have looked retrospectively at previous outturn and applied the current forecast methodology to understand the degree of error that could be expected.
Applying the forecasts in retrospect shows that the largest implied forecast error would have been +0.9 percentage point difference compared to forecast in 1997-98. The average deviation over the 20 year period, where we have data, is 0.25 percentage points for under-forecasts and for over-forecasts it is 0.35 percentage points. It therefore seems reasonable to expect that the scale of any forecast error will not exceed 1 percentage points. By way of illustration, an error of +1 percentage point (implying higher forecast than outturn) would lead to an overestimate of NDRi of around £15 million in the first year of the forecast period and £30 million thereafter.
Scottish Fiscal Commission view of the forecast
The methodology underlying the economic determinants of NDRi was assessed as reasonable by the Commission in their Final Report on the Draft Budget 2017-18. Further details on their views together with additional analysis and commentary are set out in their Final Report.