Annex A: Further detail on analysis and economic modelling
Overview of the economic modelling
This Annex describes the approach to estimating the impact of different income tax policy proposals.
Estimating the direct (or “static”) impact of policies on revenues
The starting point for forecasting income tax - both under the status quo and under any policy proposal - is to project forward Scottish income tax revenue. This is done by combining data on Scotland’s current income distribution  and income tax receipts with forecasts of growth in the number of taxpayers and their income.
The number of taxpayers is projected separately for different age bands to factor in demographic shifts in the Scottish population. This forecast is informed by the 2014-based Office for National Statistics ( ONS) principal population projections, labour market data on participation and unemployment in Scotland and projections of future employment growth.  The growth rates of different types of income, such as income from employment, pensions and property income, are also forecast separately. These forecasts take into account outturn data from the Survey of Personal Incomes and Annual Survey of Hours and Earnings where available. For simplicity, forecasts for earnings and employment growth are based on the OBR’s latest Economic and Fiscal Outlook from March 2017.  Further detail on the forecasting methodology can be found in the report published alongside the Draft Budget 2017-18. 
By comparing the forecasts under each policy proposal with those under the current tax regime, we can estimate how much additional tax a policy proposal would raise. These estimates are also referred to as “static” or “direct” policy costings.
Estimating taxpayers’ behavioural responses
Taxpayers will change their behaviour in response to a change in income tax, for example by changing the number of hours they work.
The scale of the behavioural response can be estimated by using taxable income elasticities ( TIEs). These measure the percentage change in taxable income in response to a one per cent change in the net of tax rate: the higher the TIE, the greater is the behavioural response and the greater the resulting impact on tax receipts. To ensure consistency, the same set of TIEs was applied across all policy options. These were informed by discussions with a number of institutions and a review of the academic literature. 
While there is a general consensus in the academic literature that TIEs are much higher for those on higher incomes, the exact magnitude of the behavioural response is uncertain. Therefore, the analysis assumes that TIEs are close to 0 for taxpayers on low incomes and increase as incomes rise. To highlight the degree of uncertainty, the analysis adopts a range of 0.35 to 0.75 for the TIEs of the very top earners, i.e. those earning more than £250,000, and slightly smaller ranges for taxpayers earning between £150,000 and £250,000. Since behavioural responses are much smaller at the lower end of the income range, a single TIE was applied to taxpayers earning up to £150,000.
These TIEs have been updated compared to what was set out in the Devolved Taxes Methodology Report to reflect recent research by the Institute for Fiscal Studies ( IFS), which found that those just above the £150,000 Additional Rate threshold appear to be less responsive to tax changes than the Additional Rate taxpayer group as a whole.  Likewise, the International Monetary Fund ( IMF) recently published new analysis showing that there is no evidence of an increase in income tax elasticities for top earners over time, despite cheaper access to international tax planning or increased mobility of taxpayers. 
All TIEs are kept under regular review as new information on the behaviour of Scottish taxpayers becomes available. It is important to remember that the SFC may reach different views on taxpayer responsiveness as the responsibility for producing the final income tax forecasts now sits with the SFC and its Commissioners. The Scottish Government’s assessment of behavioural responses continues to be informed by the advice by the Council of Economic Advisers on the revenue risks for raising the AR to 50p. Further advice is expected ahead of the Draft Budget.
Estimating impact on household income
As well as modelling the impact of a tax proposal on each individual’s earnings, the impact on household income is also calculated. This is because household income differs from individual earnings since households may have more than one earner and more than one source of income. The analysis therefore calculates the impact of proposals on household incomes after tax and social security payments, accounting for the number of people in the household.
Household income distributional analysis is based on a static tax and benefit micro-simulation model developed by the Scottish Government. The model is based on 2015-16 Family Resources Survey ( FRS) data.  It does not incorporate the impact of any behavioural response or secondary impacts, for example, employment impacts from the interaction between employment and social security income.
In order to replicate the impact of the introduction of a policy in 2018-19, income tax parameters for each proposal were adjusted to 2015-16 levels.  In order for comparisons to be made with incomes in that year, the most recent for which data are available. The distributional impact of the policy is then demonstrated using equivalised household income deciles. This splits all households in Scotland into 10 equally sized groups by level of income. The 1st decile is the 10% of households with the lowest incomes, whilst the 10th decile is the 10% of households with the highest incomes. Equivalised income is used as it adjusts for the composition of different households.
Further detail of the impact of behavioural responses on revenue
Table A1: Comparison Of Revenue Implications With And Without Behavioural Responses By Proposal/ Approach (£Millions In 2018-19)
|Revenue implications (no behavioural response)||Revenue implications (post-behavioural response) Mid point||Revenue implications (post-behavioural response) Low||Revenue implications (post-behavioural response) High|
|Scottish Liberal Democrats||500||444||451||437|
|Alternative approach 1||130||82||89||75|
|Alternative approach 2A||330||230||252||208|
|Alternative approach 2B||390||232||271||194|
|Alternative approach 3||430||255||294||217|
|Alternative approach 4||350||186||224||148|