Fairer Council Tax: consultation analysis

Introduction

Background and context to the consultation

Councils provide a wide range of vital services that communities rely on. This includes, but is not limited to, funding for schools, social care, roads and transport, as well as environmental and waste services. The provision of these essential services is partly made possible by the revenue generated from Council Tax, which all households pay unless exempt. In 2023, Council Tax is projected to contribute approximately £2.9 billion to funding such services in Scotland.

Council Tax has been devolved to the Scottish Parliament since 1999, but the administration of the system is the responsibility of individual councils. The tax is paid by the occupiers of domestic properties, and the amount depends on the valuation band that a property is in (A-H). A valuation band reflects an Assessor’s assessment of the open market value of the property if sold at 1 April 1991: for example, a property assesed to have an open market value of up to £27,000 would be classified in Band A.^[3]

Each council sets the tax rate for Band D properties, with the remaining bands being proportions (known as multipliers) of this charge. The multipliers are set in law and therefore the same for all councils. The amount that an individual pays will depend on:

• The valuation band the property is in,
• The Band D tax rate set by the local authority,
• Any discounts (such as the 25% single person discount), and
• Any exemptions or reductions (e.g., the Council Tax Reduction Scheme).

While the importance of Council Tax and the services and support that it helps fund are widely accepted, the structure and functioning of the present Council Tax system have received criticism from some stakeholders. Although the most financially vulnerable people in Scotland are protected through the Council Tax Reduction Scheme^[4], under the current system the effective Council Tax rate is higher for lower value properties (when expressed as a percentage of the estimated property value). This means that those residing in less expensive properties are paying a higher proportion of their property’s value in tax than those in more expensive homes, and the system has therefore been considered unfair and regressive.

This discrepancy was partly addressed in 2017 when the Scottish Government increased the Council Tax Band E to H multipliers, meaning an increase in Council Tax for those living in higher value properties, however the current system is still regressive with respect to property value.^[5] One approach to address unfairness in the system would be to repeat the changes made in 2017, and thereby increase the Council Tax charges by 7.5%, 12.5%, 17.5%, and 22.5% for Bands E through H respectively, potentially raising an additional £176 million with around 28% of all properties in Scotland impacted. While this proposal would not fully resolve the issue, such a change could play a key role in making council tax less regressive when considering the average Council Tax liability as a proportion of property value.

Through the Joint Working Group on Sources of Local Government Funding and Council Tax Reform (JWG), the Scottish Government and Convention of Scottish Local Authorities (COSLA) sought the public’s views on this proposed change to the current system. A public consultation was open from 12 July until 20 September 2023, and aimed to collect views and opinions on:

• Whether households in properties in the highest Council Tax bands (Bands E to H) should make a greater contribution, and
• The relationship between the valuation band a property is in, and the tax rate set by the council for the local authority area.

Consultation format and structure

The joint Scottish Government and COSLA consultation on a fairer Council Tax was hosted online on the Scottish Government’s Citizen Space portal. It consisted of nine questions (three closed-format questions, three open-format questions and three closed-format questions which asked respondents to give reasons for their answer in a free text field). The Scottish Government also accepted responses provided via e-mail or post. The consultation was open for 10 weeks, with 15,628 responses received in total. A full list of consultation questions is provided in Annex A.

Alma Economics, an independent research consultancy, was commissioned to analyse the responses to this consultation. This report provides a summary of the analysis in relation to the questions the Scottish Government asked during the consultation.

Data processing and cleaning

At the start of the consultation analysis, responses from Citizen Space and those sent by e-mail were merged into a single anonymised and final dataset. Responses received by e-mail were reviewed, entered manually into the final dataset and analysed alongside responses submitted through the online platform. Five responses received by e-mail or post were unstructured and did not follow the prescribed question format. Where possible, the research team mapped these responses to specific consultation questions. As the responses did not clearly indicate specific answer choices, these responses were not reflected in the breakdowns of the totals of quantitative analysis.

All responses were treated equally regardless of how they were submitted. During the manual review of responses, the research team screened for those that were clearly intended as offensive, abusive or explicitly vulgar, with no responses being removed as a result of this screening.

The consultation responses were also screened to identify duplicate responses or campaigns organised by external groups or individual coordinated responses to the consultation. Campaign responses were defined as responses which shared at least two exact duplicate sentences. Across all six open-format questions, the research team found that no more than 0.2% of responses could be classified as campaign responses, and no more than 2% of responses could be classified as exact duplicates (once very short responses that did not answer the consultation question were removed).^[6]

Respondent type

Respondents were asked whether they were responding to the consultation as an individual or organisation. These responses were self-reported and not independently verified by the research team. Two respondents were reclassified from organisations to individuals, as their response to the input field “Full name or organisation’s name” made it clear the respondents were not organisations.

Council tax band

Respondents were asked to indicate which Council Tax band(s) apply to the property (or properties) for which they pay Council Tax. These Council Tax bands were self-reported and were not independently verified by the research team. Responses to closed-format questions in the main body of the report are broken down by respondents in Council Tax bands A-D (who would not be impacted by the proposed increase) and bands E-H (who would be impacted by the increase). This breakdown did not include respondents which selected multiple tax bands, “I don’t know” or “I don’t pay Council Tax” as responses or did not answer the question (these individuals and organisations only made up 6% of total responses to the consultation). A breakdown of responses to closed-format questions by individual tax bands is included in Annex C of the report.

Approach to analysis of closed-format questions

Descriptive analysis was conducted on the responses to the six closed-format questions using Python. For each question, the research team calculated both the proportion of all respondents and proportion of subgroups of respondents (based on respondent characteristics) selecting each answer choice. The main body of this report presents the overall findings plus a breakdown of responses to each consultation question by Council Tax bands A-D and E-H, with additional insights presented for breakdowns by respondent type, individual Council Tax band and local authority. Each question includes a chart that summarises responses as a percentage of all consultation respondents who answered the question by selecting one of the available options, including “Don’t know”. Full tables of results broken down by all respondent characteristics are presented in Annex C.

An earlier version of the Respondent Information Form (before the consultation was launched on Citizen Space) included the additional question “The proportion of properties in each valuation bands E-H varies across the country. This means that some councils would benefit more than others from any increases in council tax in these property bands. Should steps be taken to ensure that all councils benefit proportionately from this policy?” (closed question) and “Please give reasons for your answer?” (open question). As this question was removed from the online version published on Citizen Space, only 43 respondents who submitted responses via e-mail answered this question. Because the number of respondents to this question was significantly smaller than the total number of respondents to the consultation (0.3%), the research team has not included any analysis of responses to this question in our report.

Approach to analysis of open-format questions

The consultation included six questions with free-text fields, and there was no limit to the amount of text which respondents could write in their answers. The research team followed an approach to analyse all responses to the consultation that combined manual and automated coding due to the substantial amount of data to review (some responses to a single question were over 2,000 words long). Broadly, the approach taken to analyse the qualitative responses comprised three steps (more detail on this approach is set out in Annex B):

1. Initial manual coding of a sample of 250 responses for each question: This sample was representative in terms of Council Tax band and local authority and weighted towards longer responses. All manually coded responses were double-checked by a second researcher, then codes were added and organised in a separate codebook.

2. Automated text analysis of responses not manually coded, using a set of Natural Language Processing (NLP) models: Outputs from the models were manually reviewed by the research team, with feedback passed to the models for further iterations. This continued until the researcher agreed the codes assigned by the automated text analysis aligned with codes that would have been manually assigned by the researcher.

3. Thematic analysis: For each open-ended question, a descriptive summary has been presented of key themes emerging from the integrated manual and automated text analysis. While it is difficult to provide accurate counts of responses allocated to each theme, in general themes are presented in the report in approximate order of the number of corresponding responses.

Individual quotes have been included where appropriate to illustrate the narrative around specific themes, and quotes were only selected from respondents who provided permission for their views to be published and with any potential identifiers (such as the name of a specific organisation) removed. Quotes were corrected in cases of identifiable personal information, typos, and missing punctuation.

Limitations

• Any information cited in quotes may not be accurate and has not been validated or fact checked. Instead, the quotes represent, and should be treated as, the subjective views or experiences of respondents alone, rather than as factual information.
• The responses to the consultation from individuals and organisations are in general unlikely to be representative of the broader population, of any specific demographic group or to be representative of geographical areas such as local authorities. Responses from some areas or groups may be disproportionately under- or over-represnted: for example, respondents from bands E-H were much more likely to respond to the consultation than those from bands A-D.
• The research team could not verify whether all responses identifying as organisations were submitted in an official capacity.
• Some responses did not directly address or relate to the questions set out in the consultation (which are the focus for this report including the qualitative analysis sections).
• All three text classification methods rely on probability-based assignment of themes to topics. Therefore, the analysis is not comparable to full manual coding of the entire dataset (particularly for themes with only a small number of example responses), and it is not appropriate to interpret the outputs with reference to precise counts or measures of accuracy such as confidence intervals.

Structure of the report

The report initially sets out details about the sample of respondents to the consultation. This is followed by sections which cover each of the consultation questions. Each section contains the findings from the quantitative analysis, followed by the qualitative analysis of respondents’ free-text answers. Many themes that were discussed in responses to the open-format section of Question 1 were also referenced in responses to the five other open-format consultation questions. To keep repetition between sections to a minimum, these themes are discussed in detail in the section for Question 1, and responses to other questions discussing these same themes are summarised briefly in the corresponding remaining section of the report.