Examining Relations Between Public Participation and Public Expenditure: Opinions from English and French Users on Environmental Issues in the English Channel

: Governments need to decide how to allocate their public expenditure, which is commonly misconstrued as simply targeting social issues. Most scientiﬁc literature highlights that the role of public spending is to enhance social welfare and ﬁght poverty and inequality. Nonetheless, public expenditure also includes spending on environmental issues. This paper analyses relations between public participation, support for public expenditure, and pro-environmental behaviour (PEB) intentions in the English Channel region. An online public survey was developed to investigate public use of the English and French sides and the public’s willingness to change their behaviour to better protect the Channel region. The survey was undertaken in the summer of 2014 and was answered by 2000 respondents. The Channel region public is willing to participate more in behaviour that involves direct changes or switches between buying / purchasing choices. In contrast, there is less willingness to engage in pro-environmental behaviour intentions that involve more active engagement activities. French respondents were slightly less inclined to change their consumer behaviour intentions, while women and older people were slightly more likely to do so. This research shows that pro-environmental behaviour could positively a ﬀ ect support for proposed public expenditure on environmental issues.


Introduction
Governments have to decide how to allocate their expenditure [1] since budgets are limited [2]. Public spending can effect growth and distribution [3] and determine regional development [4]. The level and composition of expenditure can be used to influence policy objectives [3].
Most scientific literature highlights that the role of public spending is to enhance social welfare and fight poverty and inequality [2,5], and there is a popular misconception that public expenditure refers solely to social welfare programmes [2]. Nonetheless, public expenditure also addresses issues including crime prevention, defence, science, technology and public education [5], as well as important spending on environmental issues [6].
level includes a study of public opinion on the key issues facing the European Union (EU), with issues including immigration, the economy, and environment [20]. Understanding the public's views on future priorities for the governance of the environment can enable national and local government authorities to make informed decisions regarding future funding priorities and management, and lead to improved cooperation between stakeholders, institutions and governments [21].
Various environmental problems are rooted in human behaviour, which needs to be changed, since the exclusive use of technical solutions tends to be insufficient [15]. Pro-environmental behaviour (PEB) is an essential part of orienting societies towards a more sustainable future [22]. Environmental behaviour is determined by a combination of situational, psychological and value-based factors that provide a complex response by citizens [23]. According to Stern et al. [18], policy support is influenced by pro-environmental personal norms, which are affected by personal values. For example, higher levels of pro-environmental behaviour are more likely to result in reduced meat consumption [24], with self-interested motives and pro-social motives playing significant roles in an individual's intention formation when deciding on choosing organic menu items when dining out [25], norms, values and beliefs being associated with travel mode choice [26], and willingness to address climate change issues being positively correlated among all types of climate-friendly actions [27]. Perceived behavioural control, attitudes, and moral norms are the strongest predictors of pro-environmental intentions and behaviours [28]. The theory of planned behaviour (TPB) highlights that intentions are the strongest predictor of future behaviour [29].
Most research on public participation and the environment has been related to climate change issues (e.g., [22,25]), to anthropogenic impacts on the marine environment [30,31], or to flooding and sea-level rise [32]. However, there has been limited research on public participation and the use of marine and coastal environments [33]. At the regional level, there has been research into public preferences for use of the Baltic Sea [34]. At the national level, in the United Kingdom (UK) there have been studies performed on public engagement with, and attitudes towards, the wider environment [35], attitudes towards marine protection and the marine environment [21,36,37], and on public participation in making local environmental decisions [38]. There has been little research on linking public participation, public expenditure, and pro-environmental behaviour, especially for marine environments.

Methods
An online public survey was developed to investigate public use of the English Channel as a leisure resource, the public's preferences for spending public money on the region in general and on the marine and coastal environments more specifically, and the public's willingness to change their behaviour to better protect the Channel region. The survey was part of a research study conducted from 2014 to 2015. All questions were closed-ended.
An online survey was undertaken in the English Channel region in the summer of 2014. The survey was funded by the Interreg Europe programme (Interreg, undated) which provides funding for inter-regional cooperation projects under the Promoting Effective Governance of the Channel Ecosystem (PEGASEAS) Project, which has 14 participating organisations, including academic and local government agencies (seven organisations in each of England and France). The survey was conducted by Global Marketing Insite (GMI); GMI changed its name to GMI Lightspeed subsequent to the carrying out of the survey. GMI holds information on country, gender, age, employment, and education for the panel of potential survey respondents globally.
The survey questions were developed to combine both the requirements of Interreg for data on public funding preferences in the Channel region, and the research agendas of academic partners in the PEGASEAS Project. The survey covered the areas of southern England and northern France, as illustrated in Figure 1. The responses to the online survey were received from all the English counties and French départements identified in Figure 1, since all are located in the France (Manche)-England region, as defined under the Interreg V programme for 2014 to 2020. The survey had four sections, the first of which covered basic information such as where the respondents lived (selected from the list of Interreg eligible areas as set out in Figure 1), the type of area they lived in (urban, suburban, village/rural or other) and their employment status (for example, in full time employment, self-employed, or retired).
The second section asked how frequently the respondents visited the English Channel region (in France, England or on both sides of the Channel), why they visited the Channel Region (for a holiday, work, recreation, to live there, for travel or another reason) and what they did when they visited the region (respondents could select as many options as were applicable from 15 types of activity).
The third section asked respondents to rank a number of funding priorities for the English Channel that had been identified by the funding body, Interreg, and used a five-point scale from not important to very important. From the online survey Interreg sought information under the broad themes of business and local economy, renewable energy, tourism and natural and cultural heritage, environment, and regeneration and deprivation. The respondents were asked to rate 13 specific priorities that could be funded to improve the English Channel coastal region in order to help direct the Interreg funding agenda for the period 2014-2020. The Interreg funding priorities used are set out in Table 1. Additionally, in the third section, the respondents were provided with a list of 17 marine and coastal environment-specific funding preferences, and were asked to select their five most favoured and five least favoured preferences. These funding preferences were as follows: protecting plants and animals in the sea; protecting plants and animals on the coast; working with businesses to become more sustainable and eco-friendly; creating new job opportunities on the coast and in the seas; promoting marine recreation and leisure opportunities; supporting the fishing industry; encouraging eco-friendly The survey had four sections, the first of which covered basic information such as where the respondents lived (selected from the list of Interreg eligible areas as set out in Figure 1), the type of area they lived in (urban, suburban, village/rural or other) and their employment status (for example, in full time employment, self-employed, or retired).
The second section asked how frequently the respondents visited the English Channel region (in France, England or on both sides of the Channel), why they visited the Channel Region (for a holiday, work, recreation, to live there, for travel or another reason) and what they did when they visited the region (respondents could select as many options as were applicable from 15 types of activity).
The third section asked respondents to rank a number of funding priorities for the English Channel that had been identified by the funding body, Interreg, and used a five-point scale from not important to very important. From the online survey Interreg sought information under the broad themes of business and local economy, renewable energy, tourism and natural and cultural heritage, environment, and regeneration and deprivation. The respondents were asked to rate 13 specific priorities that could be funded to improve the English Channel coastal region in order to help direct the Interreg funding agenda for the period 2014-2020. The Interreg funding priorities used are set out in Table 1. Additionally, in the third section, the respondents were provided with a list of 17 marine and coastal environment-specific funding preferences, and were asked to select their five most favoured and five least favoured preferences. These funding preferences were as follows: protecting plants and animals in the sea; protecting plants and animals on the coast; working with businesses to become more sustainable and eco-friendly; creating new job opportunities on the coast and in the seas; promoting marine recreation and leisure opportunities; supporting the fishing industry; encouraging eco-friendly developments around ports; encouraging offshore marine renewable energy; enhancing safety at sea; promoting marine pollution prevention; improving coastal flood defences; identifying priorities for coastal adaptation to climate change; ensuring clean water and beaches; creating stronger cultural links across the Channel; promoting cultural heritage and the arts around the Channel; developing better transport links across the Channel; and promoting research to support the better management of the Channel. The respondents were requested to rank the level of importance they placed on each funding priority, using a scale from not important to very important. These preferences have been considered elsewhere [21] and are not considered in this paper. Table 1. Public priorities for the INTERREG V-A France (Channel)-England cross-border cooperation programme 2014-2020.

Theme
Public Priority

Business and local economy
To support and develop future sustainability in business To help businesses better respond to economic pressures and/or create new jobs To strengthen and build networks between businesses and other stakeholder groups

Renewable energy
To further research into renewable energy technology and its potential impacts (on land and sea) To increase the use and awareness of renewable energy by businesses and the public Tourism and natural and cultural heritage To promote tourism and interest in the history, culture and geology and other attractions on the Channel coast To support local businesses providing services or goods to visitors and tourists to the Channel Coast

Environment
To raise public awareness of the Channel environment (e.g., through campaigns and social media) To reduce pollution and improve the management of environmental risks To improve the management of natural resources and conservation of the Channel environment To increase awareness of the benefits that the Channel environment provides to humans (e.g., fish, leisure and recreation, and health) To support adaptation to climate change Regeneration and deprivation To support physical, economic and social regeneration in deprived urban and rural communities The fourth section of the survey examined PEBs of the respondents. Academic partners within the PEGASEAS Project developed a list of PEBs and the survey asked respondents, based on their knowledge and previous responses to the survey, to identify if they had, or would be willing to, change their behaviour to protect the environment. They were provided with eight options with respect to changing their lifestyle (including whether they could, could not, or already had changed their behaviour), and with 11 types of behaviour (from buying sustainably sourced fish to participating in marine planning activities). These pro-environmental behaviours and options for change are set out in Table 2. The survey was initially tested by 200 respondents in total divided equally between England and France, as questions were provided in their native language in order to ensure that they were clear and not open to misunderstanding. As there were no changes needed to the survey, it was subsequently sent to more participants, until 999 responses from England and 1001 from France were received. The survey required all respondents to be over the age of 16 and reside in one of the eligible areas covered by the Interreg V programme for the Channel region, a list of which was provided to them. A breakdown of survey respondents is provided in Table 3.

Data Analysis Methods
Non-parametric methods were used because many of the variables (for example PEBs) were measured in ordinal scales. The statistical analyses were conducted using IBM SPSS 22 for Windows [41]. A research model was developed to analyse the data (see Figure 2). A quantitative analysis was performed using the following techniques: first, principal component analysis (PCA) and non-linear PCA (NLPCA) were used in order to assist further analysis. An NLPCA is similar to an ordinary PCA but it can be applied to variables that are not ratio or interval scales, such as the ordinal scales used in this survey [42]. NLPCA uses a process called quantification to replace the original values with optimally scaled ones and then conducts a PCA. The number of variables was reduced using this technique. Second, several multiple linear regressions were performed. As a first stage, a regression analysis was undertaken looking at respondent characteristics and behavioural intentions (rows 6 to 18 in Table 4). A regression analysis was also undertaken looking at respondent characteristics and willingness to spend public funding (rows 19 to 29 in Table 4). A regression analysis was then undertaken between behavioural intentions and willingness to spend public funding. Table 4 shows the main descriptive data (minimum, maximum, mean and standard deviation) of the items. Within the behavioural variables (rows 19 to 29 in Table 4) there were many missing values. These were, therefore, replaced by mean values in the regression analysis.
technique. Second, several multiple linear regressions were performed. As a first stage, a regression analysis was undertaken looking at respondent characteristics and behavioural intentions (rows 6 to 18 in Table 4). A regression analysis was also undertaken looking at respondent characteristics and willingness to spend public funding (rows 19 to 29 in Table 4). A regression analysis was then undertaken between behavioural intentions and willingness to spend public funding. Table 4 shows the main descriptive data (minimum, maximum, mean and standard deviation) of the items. Within the behavioural variables (rows 19 to 29 in Table 4) there were many missing values. These were, therefore, replaced by mean values in the regression analysis.

Limitations of the Methods
Online surveys have limitations such as self-selection bias [43] and sample representativeness [44], as well as some possible small inherent bias from sampling respondents registered on a database with a market research company [43]. The difference in education levels between the groups may reflect differences in the education systems of the two countries. Only respondents residing close to the English Channel were included, which limits the findings to the surveyed population. Although beyond the scope of this study, including other comparable communities in other regions and countries would have increased generalisability. The data was collected as part of the PEGASEAS project. In ideal circumstances, data should have been collected at the beginning and at the end, but this was not possible due to the difficulty of tracing the respondents, combined with their anonymity. An adequate level of R 2 is not universally accepted, with some authors assigning different thresholds depending on the characteristics and criteria of their respective studies [45], whereas others stating that it is inappropriate to assign a specific R 2 threshold [46]. It is important to observe that residuals behave randomly to assess a model fit [47]. However, a relatively low R 2 value suggest that factors, other than those included in a regression model, explain a larger portion of the variation of the dependent variable. Table 4 shows the results of the survey, where it can be seen that the variables with the highest mean (after normalisation against each variable's maximum) were: use of fewer plastic bags (PEB category); improved management of natural resources and conservation of the Channel environment (environment category); reduced pollution and improved management of environmental risks (environment category); promotion of tourism and interest in local history, culture and geology and other attractions on the Channel coast (tourism and natural and cultural heritage category); increased awareness of the benefits that the Channel environment provides to humans (environment category); support for local businesses providing services or goods to visitors and tourists to the Channel Coast (tourism and natural and Sustainability 2019, 11, 2230 9 of 14 cultural heritage category); further research into renewable energy technology and its potential impacts (renewable energy category); support for and development of future sustainability in businesses (business and local economy category); help for businesses to better respond to economic pressures and/or create new jobs (business and local economy category); support for physical, economic and social regeneration in deprived urban and rural communities (regeneration and deprivation category); and support adaptation to climate change (environment category).

Reducing Number of Dimensions with Principal Component Analysis and Non-linear Principal Component Analysis
The constructs 'behavioural intentions' and 'willingness to support public spending' consisted of 11 and 13 items, respectively. In order to assist with further analysis, the number of variables used was reduced using PCA. The items on behavioural intentions were measured using ordinal scales, which meant that NLPCA was more appropriate than ordinary principal component analysis [42]. The variables were set as ordinals and the ranking method was used for discretization.
The outcomes of the PCA and the NLPCA were rotated using Promax (Tables 5 and 6). For the behavioural intentions dimension (see Table 5), the NLPCA produced two coherent dimensions which will be called 'participation' and 'consumer behaviour' in subsequent analysis. Both dimensions show satisfactory loading (the two columns in Table 5) and Cronbach alpha values. For the 'willingness to support for public expenditure' construct (Table 6), two items had to be removed in order to get two coherent dimensions. These were 'to raise public awareness of the Channel environment, e.g., through campaigns and social media' and 'to increase awareness of the benefits that the Channel environment provides to humans (e.g., fish, leisure and recreation, and health)'. After the two items were removed, the two dimensions did indeed have satisfactory component loadings and Cronbach alpha values. The two dimensions are named 'economic' and 'environmental' in subsequent analysis, reflecting the support for spending public money within these two areas.

Component Economic Environmental
To support and develop future sustainability in businesses 0.694 To help businesses better respond to economic pressures and/or create new jobs 0.831 To strengthen and build networks between businesses and other stakeholder groups 0.773 To further research into renewable energy technology and its potential impacts (on land and sea) 0.797 To increase the awareness and use of renewable energy by businesses and the public 0.751 To promote tourism and interest in history, culture and geology and other attractions on the Channel Coast 0.720 To support local businesses providing services or goods to visitors and tourists to the Channel Coast 0.879 To reduce pollution and improve the management of environmental risks 0.874 To improve management of natural resources and conservation of the Channel environment 0.619 To support physical, economic and social regeneration in deprived urban and rural communities 0.674 To support adaptation to climate change (e.g., environmental management and research) 0

Impact of Respondent Characteristics on Behavioural Intentions
In order to test the research model (Figure 2), multiple linear regression was used. In the first two regression models (Table 7), the impact of the characteristics of the respondents on behavioural intentions was tested. Five independents variables were used, of which four were dummy variables: country (France = 1), gender (female = 1), employment (in employment = 1) and education (university education = 1). The fifth variable was the age of the respondent. The results in Table 7 show that the independent variables' impact on behavioural intentions is generally rather low, although still significant for country, gender and age. The residuals for the participation model are skewed, which shows that the overall model is not satisfactory and the result are inconclusive. It cannot be determined whether any of the five items have any impact on 'participation' behavioural intentions. For 'consumerism' the residuals do indeed behave randomly, which strengthens the validity of the model. The results show that the French are slightly less inclined to change their consumer behaviour, while women and older people are slightly more likely to do so. Employment and education do not have any impact on consumer behavioural intentions. Note that the R 2 values are rather low, implying that other factors are needed in order fully explain behavioural intentions.

Impact of Environmental Intentions on the Willingness to Support Spending Public Money (While Controlling for Respondent Characteristics)
The second stage of regression models (Table 8) tested the impact of respondent characteristics and behavioural intentions on support for spending public money. Multiple regression analyses were utilised, introducing clusters of variables sequentially. First, the impact of respondent characteristics was tested. Second, the 'participation' variable was added, and, then, the 'consumer behaviour intention' variable was included in the last multiple regression analyses. The results show that country, gender and age have some impact on willingness to support spending public money: the French are more likely to support public spending, as are women and older people. This partially contradicts the findings of Park [13] and Svallfors [14], who indicated large differences in attitudes between countries. The explanatory power of the first model is low. The impact of respondent characteristics is much lower than the impact of the two types of behavioural intentions: 'participation' and 'consumer behaviour'. These 'behavioural intentions' are significantly associated with willingness to support spending public money, and are the strongest predictors for future behaviour (see [29]). The results imply that people who intend to be more engaged in 'participation' are significantly more likely to support public spending and are particularly keen on public spending on ecology. People who intend to change their consumer behaviour are also likely to support public spending and are also keener on spending on ecology rather than on economic development. It should be noted that 'participation' has a much stronger impact on support for public spending than 'consumer behaviour'.
The explanatory power (R 2 ) is not particularly high for any of the models, although it is higher for spending on ecology than on spending on the economy. However, the F value is significant, and the residuals behave randomly, indicating the data fits the model well. The model shows that the variables included have an impact on willingness to spend public money but only explain a limited part of the variation. Other variables, not considered in this research, may have an equal or even bigger impact on willingness to spend public money.

Conclusions
Governments have to decide how to allocate their public spending, which is commonly misinterpreted as covering solely social issues (such as fighting poverty and inequality, welfare issues, and social programmes). However, public spending entails other social issues, as well as, importantly, environmental ones. This paper has examined relations between public participation through pro-environmental behaviour and support for public spending on economic and environmental issues in the English Channel region. This could be considered an antecedent to full consultation on participatory budgeting, a stage prior to 'formulation' (complementing Heimans's [8] stages).
An online public survey was developed to investigate public use of the English Channel and the marine and coastal environment more specifically, as well as the public's willingness to change their behaviour to better protect the Channel region. The survey was undertaken in the English Channel region in the summer of 2014 and was answered by 2000 people in total in France and England.
A positive impact of environmental intentions on willingness to support spending public money was found while controlling for respondent characteristics. It was found that the Channel public participate more in behaviours that involve direct changes or switches between buying/purchasing choices. In contrast, there is less participation in PEBs that involve more active engagement activities, for example, meetings, groups, campaigns and politics. This research provides a comprehensive perspective on French and English public use of environmental goods and services.
This research highlights that pro-environmental behaviour and willingness to change could positively affect participation on public spending on environmental issues. Raising awareness about the importance of change in consumer behaviour could be achieved through general training in environmental issues, and, in this way, more support for environmental public expenditure could be achieved.
Further research should be carried out in other contexts and by studying other comparable communities in other regions and countries to explore further the relationship between PEBs, willingness to change, and public spending for environmental issues. Another topic that could be explored is the link between public policy and public spending in the context of cross-country environmental issues, as well as controlling for spending in general.