ATTITUDES TOWARDS INCOME AND WEALTH INEQUALITY AND SUPPORT FOR SCOTTISH INDEPENDENCE OVER TIME AND THE INTERACTION WITH NATIONAL IDENTITY

Transcription

1 Scottish Affairs 23.1 (2014): DOI: /scot # Edinburgh University Press ATTITUDES TOWARDS INCOME AND WEALTH INEQUALITY AND SUPPORT FOR SCOTTISH INDEPENDENCE OVER TIME AND THE INTERACTION WITH NATIONAL IDENTITY Claire L. Niedzwiedz and Mor Kandlik-Eltanani Abstract As the Scottish independence referendum is drawing near, the importance of understanding public attitudes that influence support for independence is increasing. The relationship between attitudes towards social inequality and support for independence is not well understood. Using data from the Scottish Social Attitudes Survey, we empirically analysed the relationship between attitudes towards income and wealth inequality, taxation, and income redistribution and support for independence over time. Further, we examined to what extent party identification explained the findings and the interaction between attitudes towards social inequality and national identity. Individuals who held more left-wing attitudes towards social inequality were more likely to support independence and the associations were stronger among individuals identifying as more Scottish. Over time, the associations Claire Niedzwiedz is a doctoral researcher within the Institute of Health and Wellbeing at the University of Glasgow and is interested in the social and structural determinants of health and health inequalities. Claire L. Niedzwiedz, Institute of Health and Wellbeing, 1 Lilybank Gardens, University of Glasgow, Glasgow, G12 8RZ. claire.niedzwiedz@gmail.com Mor Kandlik Eltanani is a doctoral researcher within the School of Social and Political Science at the University of Edinburgh and focuses on research of social movements, as well as quantitative research and teaching. Mor Kandlik-Eltanani, School of Social and Political Science, University of Edinburgh, Edinburgh, EH8 9LD. Mor.kandlik@gmail.com 27

2 Claire L. Niedzwiedz and Mor Kandlik-Eltanani remained relatively consistent. Party identification explained some, but not all of the associations. There was evidence to suggest that having a positive attitude towards the government s role in income redistribution may be becoming more important for independence support in recent years, especially among the more Scottish groups. The opposite was found for general attitudes towards wealth inequality. Further analysis is required to investigate whether these trends continue as the campaigns for and against independence become more active. Keywords: Scottish independence; national identity; survey analysis; inequality; redistribution; referendum. Introduction The decision to support Scottish independence is influenced by a variety of factors including an individual s sense of national identity, socio-demographic makeup, expectations for an independent Scotland, and political ideology (Bond and Rosie 2002; Curtice 2011; Curtice 2013). Men, younger age groups, and those identifying themselves as Scottish and not British on the Moreno national identity scale (Moreno 2006) are more likely to support independence (Curtice 2013). However, less is known about how an individual s political values, particularly those relating to social inequality, are related to support for independence, and how these relationships differ by national identity. The relationship between attitudes towards social inequality and support for independence is of interest as social justice is a key focus of the Yes Scotland campaign, which proposes that an independent Scotland would be an opportunity to create a fairer, more prosperous country (Dempsie 2013). Since the beginning of the devolution process, Scottish nationalism has been closely related to welfare policy and social justice debates (Mooney and Scott 2011, 2012; Scott and Wright 2012). In addition, it is known that support for radical social democracy is more common among individuals identifying themselves as more Scottish (Paterson 2000). However, it is unclear whether, and to what extent, more left-wing attitudes towards social inequality amongst the general public help to explain an individual s decision to support independence and how this differs by national identity. It is also of interest to investigate whether any relationship between attitudes towards social inequality and support for independence could be explained by political party identification. 28

3 Attitudes towards Inequality and Scottish Independence Over the past 30 years wealth inequality has been increasing across most developed countries. In the UK, the wealthiest 1% of the population held 10% of the country s total wealth during 2008 (Bonesmo Fredriksen 2012). Some studies have suggested that societies with larger income and wealth inequalities tend to experience more health and societal problems, including poorer mental wellbeing and higher rates of violence (Lynch et al., 2004; Wilkinson and Pickett 2010). The debate which ensued emphasised the need for further analysis of inequalities in general, and of economic inequalities, in particular (Goldthorpe 2010). Since devolution, Scottish policy has diverged from England across a number of areas, including education and health. This divergence has allowed space to be opened up for challenges to neoliberalism (Scott and Mooney 2009: 387). Currently, the Scottish Government has limited control over policy levers, such as taxation and welfare policy, which have the potential to reduce social inequalities within Scotland. In an independent Scotland, the Scottish Parliament would be fully responsible for addressing these inequalities (Scottish Government 2013). This suggests that individuals more concerned with issues of social inequality may be more likely to support independence to enable the Scottish government to have more powers to influence mechanisms to tackle these inequalities, especially given recent changes in welfare policy under the current UK Conservative-Liberal Democrat coalition government. We therefore hypothesised that individuals more concerned about social inequality may be more likely to support independence. In this paper, our aims were to first quantitatively explore the association between attitudes towards social inequality and the government s role in using particular mechanisms to address social inequality, such as income redistribution and taxation, and the tendency to support Scottish independence. Second, we explored to what extent party identification explained any association. Third, we aimed to investigate the temporal trends in attitudes towards social inequality and support for independence, and differences in these relationships by national identity. Data and methods We used data from the multiple waves of the Scottish Social Attitudes Survey in a repeat cross-sectional design (from 1999 to 2012). Our dependent variable was derived from a questionnaire item asking respondents views on Scotland s constitutional future, which was included in each year of the 29

4 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Table 1 Questionnaire item on Scotland s constitutional future Which of these statements comes closest to your view? 1. Scotland should become independent, separate from the UK and the European Union 2. Scotland should become independent, separate from the UK but part of the European Union 3. Scotland should remain part of the UK, with its own elected parliament which has some taxation powers 4. Scotland should remain part of the UK, with its own elected parliament which has no taxation powers 5. Scotland should remain part of the UK without an elected parliament survey (Table 1). This item was recoded into a binary variable distinguishing those who were supportive of independence from all others. Four independent variables were investigated, representing general attitudes towards income and wealth inequality, as well as the government s role in the redistribution of income and taxation (Table 2). We initially examined the relationship of each independent variable with support for independence in cross-tabulations (Appendix 1). The independent variables were then converted into binary groups, due to the small numbers in some categories and to aid the interpretation of results. The wealth and income inequality variables were recoded into those who strongly agreed or agreed with the statements, versus all others. The income inequality variable was recoded into those who agreed that the income gap was too large compared to all others and the taxation variable was recoded into those who thought the government should increase taxes versus all others. The choice to recode the independent variables into dichotomous variables contributes to the clarity and power of our analysis. Interpreting binary logistic regression models is clearer when comparing two groups to one another, rather than comparing three or four groups at the same time. Furthermore, some of the answer categories were chosen by a small number of people, and so using these categories separately would produce a weaker analysis. However, it comes with a certain loss of information. We conducted a complete case analysis (for each independent variable) and therefore excluded individuals with missing values on the outcome 30

5 Wealth inequality (available 2000, 2002, , ) Ordinary working people do not get their fair share of the nation s wealth 1. Agree strongly 2. Agree 3. Neither agree nor disagree 4. Disagree 5. Disagree strongly Income redistribution (available 2000, 2002, , ) Government should redistribute income from the better-off to those who are less well off 1. Agree strongly 2. Agree 3. Neither agree nor disagree 4. Disagree 5. Disagree strongly Table 2 Questionnaire items on social inequality Income inequality (available 1999, 2000, 2001, 2004, 2006, ) Thinking of income levels generally in Britain today, would you say that the gap between those with high incomes and those with low incomes is too large 2. about right 3. too small Taxation (available , 2006, ) Suppose the government had to choose between the three options on this card. Which do you think it should choose? 1. Reduce taxes and spend less on health, education and social benefits 2. Keep taxes and spending on these services at the same level as now 3. Increase taxes and spend more on health, education and social benefits

6 Claire L. Niedzwiedz and Mor Kandlik-Eltanani (N = 1,127, 6%) or control variables (N = 982, 5%), including those who did not know the answer, or refused to answer, particular questions. A total of 2,477 (13%) individuals was missing information on one or more of the independent variables. All analyses were weighted to be representative of the Scottish population and to address patterns of non-response. For the descriptive analysis examining attitudes over time, age-sex standardised rates are presented using the 2001 Scottish population as reference (General Register Office for Scotland 2011). This enables rates to be compared across years and accounts for potential changes in the Scottish population structure. Multiple logistic regression models were used to calculate the association between attitudes towards social inequality and support for independence. The statistical models were adjusted for potential confounding factors including age, gender, education level, and national identity. Our choice of control variables was first guided by theory; we considered these variables may be associated with both attitudes towards social inequality and support for independence. We also examined the statistical associations among the variables and found that most were associated with both attitudes towards social inequality and independence support, as we will explain in detail later in the paper. Age was entered into the models as a continuous variable. Education level was derived from the respondents self-reported highest educational qualification and was recoded into four categories: no qualifications; Standard grades or GCSEs; Highers or Advanced Levels; or a higher education qualification. Individuals with foreign or other qualifications were excluded from the analysis (N = 239). National identity was measured using the Moreno question (Moreno 2006). Participants were asked which best describes how they see themselves: Scottish not British, more Scottish than British, equally Scottish and British, more British than Scottish, or British not Scottish. This question is thought to measure the degree of Scottish national identity, rather than just its presence or absence (Bond 2000). The Moreno question was not included in the 2002 and 2004 survey waves, and therefore these years were excluded from the regression analysis. A total of 797 (4%) of participants was missing information on national identity, or provided a description that could not be coded reliably. We also investigated whether party identification mediated the relationship between attitudes towards income and wealth inequality and support for independence. Party identification was recorded as either: none, Conservative, Labour, Liberal Democrat, SNP, Green party, or other party. The latter two categories were combined into other due to the small numbers identifying with these groups. A total of 1,505 (8%) individuals was lacking information for this variable. A net 32

7 Attitudes towards Inequality and Scottish Independence total of 5,002 individuals was therefore missing information for one or more of the variables included in the analysis. Pooled analysis combining the data from 1999 to 2012 (excluding 2002, 2004 and 2008 as the required data were not available) was conducted to examine the overall association between attitudes towards social inequality and support for independence. These models also controlled for the survey year. As the income inequality and taxation variables were available for fewer survey years we chose to concentrate the further analysis on the wealth inequality and income redistribution variables. Interaction terms between the independent variables and the survey year were included to investigate whether there was any change in the associations over time. Next, we calculated the adjusted predicted probability of supporting independence by the two levels of the independent variable (agree versus disagree or neutral) and graphed these to aid the interpretation of results (Williams 2011). The graphs for these results show data from 2005 onwards as the wealth inequality and income redistribution questions were not asked during 1999, 2001 and 2003, and the national identity question was excluded from the 2002 and 2004 survey years. In addition, we calculated the average marginal effect of changing from 0 to 1 on our independent variables (i.e. agreeing with the statements) on the support for independence for the categories of national identity, to see if there was any evidence that the effect of more leftwing attitudes towards social inequality on support for independence varied across the national identity groups. Marginal effects aid the interpretation of logistic regression models and allow us to measure the expected change in the probability of supporting independence as a function of the agreement with statements about social inequality, while keeping the covariates constant. All analyses were performed using Stata version Results Support for independence over time The public support for Scottish independence has fluctuated since the start of devolution (Figure 1), peaking at 37% during 2005, the year of the UK general election in which the Labour party were re-elected. Support for independence has averaged at 30% over the period from 1999 to During 2012, the year after the election of the Scottish National Party as a majority government, support for independence dropped to one of its lowest levels since the start of devolution. 33

8 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Figure 1 Percentage of individuals supporting Scottish independence over time (with 95% confidence intervals) Standardised by age and sex; data not available for 2008 Attitudes towards income and wealth inequality, income redistribution, and taxation over time After looking at the fluctuations in our dependent variable, we now turn to our four independent variables. Figure 2 examines attitudes towards income and wealth inequality, income redistribution, and taxation over time. Across all years, the rate of agreement that ordinary working people do not get their fair share of the nation s wealth averaged at 63%. Since 1999 the rate of agreement has gradually fallen, but during 2007, the year the SNP were elected as a minority government, a slight increase was observed. From 2009 to 2011, following the 2008 financial crisis, the rate of agreement began to steadily increase to one of its highest levels, before decreasing again in The percentage of people who believed the income gap within the UK was too wide was generally high, averaging at 85% across the years. This variable displayed the smallest variation over time compared to other three. 34

9 Attitudes towards Inequality and Scottish Independence Figure 2 Percentage agreeing with statements about income and wealth inequality, income redistribution, and taxation over time (with 95% confidence intervals) Standardised by age and sex Apart from a slight decline in the percentage of people agreeing that the income gap was too wide from the years 2000 to 2009, it has remained relatively stable. The proportion of individuals agreeing that the government should redistribute income from people who are better-off to those who are less well-off fell steeply from 52% in 2000 to 33% in The average rate of agreement was 43% across the years examined. Since 2007, the percentage of people agreeing that the government should redistribute income has increased, to its highest level during 2011, before falling again in 2012, similar to the trend observed for the statement about wealth inequality. The proportion of people agreeing that the government should increase taxes peaked in 2001 at 66%, the year in which the Labour party was re-elected in the UK general election. Since then it has steadily declined, apart from a small increase during Across all years, the average rate of agreement that the government should increase taxes was 54%. Overall, it is interesting to note 35

10 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Table 3 Cross-tabulations of the relationship between attitudes towards income and wealth inequality support for independence (for or against) Income gap is too large Nation s wealth is unequally shared Disagree Agree Disagree Agree Against 74% 69% Against 77% 65% For 26% 31% For 23% 35% Total (N) 1,517 8,415 Total (N) 4,606 7,978 the high agreement that the income gap is too large, compared to the much lower agreement that the government should redistribute income or increase taxes. Bivariate analysis To examine the relationships between attitudes towards income and wealth inequality and support for independence, we firstly examined the bivariate relationships in cross-tabulations (Table 3). Around 74% were against independence and 26% were pro-independence among those who did not agree that the income gap within the UK was too large, compared to 69% who were against independence and 31% who were for independence among those who agreed with the statement (chi-square = 15.3, p < 0.01). A more marked difference was observed for the wealth inequality variable. Among those who disagreed that working people do not get their fair share of the nation s wealth, 77% were against independence and 23% were supportive. Among those agreeing with the statement, 35% were pro-independence and 65% were against independence (chi-square = 191.7, p < 0.01). Looking at the variables which represent mechanisms the government could use to reduce income and wealth inequality, a clearer pattern was observed between more left-wing attitudes towards the redistribution of income and support for independence compared to the taxation variable (Table 4). Among those agreeing that the government should redistribute income from those who are better-off to those who are less well-off, 37% were pro-independence and 63% were against, compared to 25% who were pro-independence and 36

11 Attitudes towards Inequality and Scottish Independence Table 4 Cross-tabulations of the relationship between attitudes towards the redistribution of income and taxation and support for independence (for or against) Government should redistribute income Government should increase taxes Disagree Agree Disagree Agree Against 75% 63% Against 71% 69% For 25% 37% For 29% 31% Total (N) 7,130 5,462 Total (N) 5,153 6,173 75% who were against independence among those who disagreed with the statement (chi-square = 192.0, p < 0.01). Among those who disagreed that the government should increase taxation, 71% were against independence and 29% were for independence, compared to 69% who were against independence and 31% who were for independence among the agreeing group (chi-square = 5.2, p = 0.02). Logistic regression models After reviewing the bivariate relationship between our dependent variable and each of our four independent variables, we conducted a series of logistic regression models. The regression models allow us to control for the effect of national identity, survey year, and several demographic variables, on our relationships of interest. Each of the independent variables had a positive association with support for independence in the pooled analysis combining the data from 1999 to 2012 in multiple logistic models (Table 5). The control variables exhibited the expected relationships; female gender, increased age, and higher educational qualifications were all associated with the decreased likelihood of supporting independence. Whereas, feeling Scottish and not British, or more Scottish than British, was associated with the increased probability of supporting independence compared to those who considered themselves equally Scottish and British. Looking at the effect of the first independent variable (Model 1 in Table 5), the odds of supporting independence were 1.58 (95% CI: 1.39 to 1.78) times 37

12 Table 5 Logistic regression models looking at the relationship between attitudes towards social inequality and support for independence Model 1 Wealth inequality Model 2 Income inequality Model 3 Taxation Model 4 Income redistribution Odds Ratio [95% CI] Odds Ratio [95% CI] Odds Ratio [95% CI] Odds Ratio [95% CI] Age 0.98*** 0.98*** 0.98*** 0.98*** [0.98,0.99] [0.98,0.98] [0.98,0.98] [0.98,0.99] Female gender a 0.76*** 0.77*** 0.75*** 0.78*** [0.68,0.85] [0.69,0.87] [0.67,0.85] [0.70,0.87] Education level b Standard grade/gcses * [0.85,1.17] [0.69,0.96] [0.74,1.02] [0.85,1.18] Highers/A levels 0.77** 0.60*** 0.57*** 0.77** [0.64,0.93] [0.49,0.73] [0.46,0.70] [0.64,0.94] Higher education/postgrad 0.79** 0.60*** 0.62*** 0.77** [0.67,0.94] [0.50,0.71] [0.52,0.73] [0.65,0.91] National identity c Scottish not British 5.76*** 5.35*** 5.73*** 5.90*** [4.89,6.78] [4.47,6.40] [4.78,6.86] [5.01,6.95]

13 More Scottish than British 2.59*** 2.58*** 2.51*** 2.56*** [2.20,3.05] [2.15,3.09] [2.09,3.02] [2.18,3.02] More British than Scottish * 0.61* 0.77 [0.51,1.04] [0.41,0.99] [0.39,0.96] [0.54,1.10] British not Scottish * [0.53,1.05] [0.45,0.99] [0.50,1.11] [0.54,1.07] Independent variables d Agree wealth unequally distributed 1.57*** [1.39,1.78] Agree income gap is too large 1.19* [1.00,1.42] Agree government should increase tax Agree government should redistribute income 1.14* [1.01,1.28] 1.65*** [1.48,1.85] Observations 9,293 8,001 7,936 9,300 CI = confidence intervals; *p < 0.05, **p < 0.01, ***p < 0.001; a Reference category is male; b reference category is no qualifications; c reference category is equally Scottish and British; d reference category is disagree or neutral; odds ratios for survey year not displayed.

14 Claire L. Niedzwiedz and Mor Kandlik-Eltanani greater for those who agreed that working people do not get their fair share of the nation s wealth compared to those who disagreed, or neither agreed or disagreed. A weaker, but positive, relationship was found for the income inequality variable (Model 2), where the odds of supporting independence were 1.19 (95% CI: 1.00 to 1.42) times larger for those who agreed that the income gap between the highest and lowest paid within the UK was too high, compared to those who thought the gap was about right or was too small. Similarly, a weak but positive association (odds ratio 1.14, 95% CI: 1.01 to 1.28) was found between support for independence and the view that the government should increase taxes and increase spending on health, education, and social benefits, compared to those who thought the government should decrease taxes (and decrease social spending), or maintain the current level (Model 3). A stronger relationship was observed for the income redistribution variable (Model 4). The odds of supporting independence among individuals who thought the government should redistribute income from the better-off to the less well-off were 1.65 (95% CI: 1.48 to 1.85) times higher compared to those who disagreed or had a neutral opinion. Introducing party identification The logistic regression models were repeated with the addition of party identification, to examine whether the relationship between our independent variables and attitudes towards independence could both be attributed to party identification. Including party identification in the models attenuated the associations between the independent variables and support for independence across all models (Table 6). Only individuals identifying themselves with the SNP were more likely to support independence, compared to those with no party identification. The relationship between the agreement that the income gap was too large and support for independence was no longer statistically significant after introducing party identification into the model. This was also the case for the association between agreement that the government should increase taxation and support for independence. However, the association between the agreement that wealth is unequally shared and that the government should redistribute income and independence support remained statistically significant. The income redistribution variable remained most strongly associated with support for independence out of the variables investigated. 40

15 Table 6 Logistic regression models looking at the relationship between attitudes towards social inequality and support for independence including party identification Model 1 Agree wealth is unequally distributed Model 2 Agree income gap is too large Model 3 Agree government should increase tax Model 4 Agree government should redistribute income Odds ratio [95% CI] Odds ratio [95% CI] Odds ratio [95% CI] Odds ratio [95% CI] Independent variable a 1.40*** *** [1.22,1.61] [0.85,1.27] [0.95,1.26] [1.39,1.80] Party identification b Conservative 0.35*** 0.30*** 0.26*** 0.35*** [0.26,0.45] [0.22,0.41] [0.19,0.35] [0.27,0.47] Labour 0.57*** 0.66*** 0.60*** 0.56*** [0.47,0.69] [0.53,0.81] [0.49,0.74] [0.46,0.68] Liberal Democrat 0.47*** 0.45*** 0.37*** 0.47*** [0.36,0.62] [0.33,0.61] [0.27,0.51] [0.35,0.62] SNP 2.66*** 2.70*** 2.43*** 2.69*** [2.18,3.27] [2.16,3.39] [1.94,3.05] [2.19,3.30] Other * 1.58* 1.36 [0.86,2.04] [1.11,2.66] [1.07,2.33] [0.89,2.10] Observations 8,225 6,698 6,640 8,225 CI = confidence intervals; *p < 0.05, **p < 0.01, ***p < 0.001; a reference category is disagree or neutral in all models; b reference category is no party identification; All models are controlled for age, gender, education level, national identity and survey year.

16 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Figure 3 Predicted probability (with 95% confidence intervals) of supporting independence over time by attitude towards wealth inequality Temporal trends Next we examined the statistical interaction of the wealth inequality and income redistribution variables with survey year to see if there was any evidence that their association with support for independence had changed over time. Figure 3 uses the results of the logistic regression models to show the predicted probability of supporting independence according to the wealth inequality variable over time, holding the control variables constant. There was no clear trend in the association over time. However, there was a decline in the probability of supporting independence from 2005 to 2007 among those both agreeing and disagreeing (or having a neutral opinion) that working people do not get their fair share of the nation s wealth. From

17 Attitudes towards Inequality and Scottish Independence to 2009, there was a steeper increase in the probability of supporting independence among those agreeing that wealth was unequally distributed, compared to those who disagreed or who neither agreed nor disagreed, perhaps related to increased awareness of wealth inequality following the financial crisis. This increased probability of supporting independence among those agreeing with the statement about wealth inequality was not sustained in 2010, however. After 2010, the probability of supporting independence fluctuated among those agreeing that wealth was unequally distributed. Among those who disagreed or had a neutral opinion there was a small increase in the probability of supporting independence. During 2012, it is interesting to note the narrowing of the gap in the probability of supporting independence between the two groups, which has been observed since Appendix 2 (Model 1) contains the full results from model testing the interaction between survey year and the wealth inequality variable. From 2005 to 2010 similar trends in the association between attitudes towards wealth inequality and support for independence were observed for the income redistribution variable (Figure 4). The gap in the probability of supporting independence between those agreeing and disagreeing that the government should redistribute income narrowed considerably in 2010, similar to that observed during 2005 and This difference then widened in 2011, mostly due to an increase in the probability of supporting independence among those agreeing that the government should redistribute income. In the latest 2012 wave of data collection, that gap remained, and although the probability of supporting independence declined in both groups it was slightly more apparent in the group that disagreed that the government should redistribute income. Appendix 2 (Model 2) contains the full results from the interaction between the income redistribution variable and survey year. Trends by national identity One of the prominent parts of the public debate around Scottish independence relates to the influence of Scottish and British national identity. Since this is an important issue in the Scottish independence debate, we examined the effect of national identity on the relationship between attitudes towards wealth inequality and income redistribution and support for Scottish independence. First, we explored the effect of national identity on the probability of supporting independence over time (Figure 5). Among the groups identifying as more Scottish than British and Scottish not British 43

18 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Figure 4 Predicted probability (with 95% confidence intervals) of supporting independence over time by attitude towards income redistribution the probability of supporting independence was much higher. The groups identifying as more British, or equally Scottish and British, generally shared similar probabilities of supporting independence. Over time, the probability of supporting independence has mostly fluctuated in the same pattern across all national identity groups. Attitudes towards social inequality and the interaction with national identity Now we turn towards the effect of attitudes towards wealth inequality on independence support by national identity. Figure 6 shows the average 44

19 Attitudes towards Inequality and Scottish Independence Figure 5 Predicted probability (with 95% confidence intervals) of supporting independence over time by national identity marginal effect of the agreement that ordinary working people do not get their fair share of the nation s wealth on the probability of supporting independence by national identity. In 2009 among the Scottish not British group, the average marginal effect of the agreement that wealth is unequally distributed was This can interpreted as an average increase of 20% in the probability of supporting independence when agreeing that wealth is unequally shared compared to disagreeing or having a neutral opinion, holding the covariates constant. Trends over time were similar for each national identity group. However, it is interesting to note the decrease from 2011 to 2012 in the effect of agreeing 45

20 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Figure 6 Effect of agreeing that ordinary working people do not get their fair share of the nation s wealth on the probability of supporting independence by national identity over time wealth is unequally shared on support for independence in all national identity groups, especially among the more Scottish than British and Scottish not British groups. In the three groups that identified with the British identity more than or equal to the Scottish identity, the effect of agreement with the wealth inequality statement on support for Scottish independence was low throughout the survey years. However, for the two groups who felt more Scottish, the effect of agreeing that wealth was unequally distributed on support for independence increased after 2005, peaked in 2009, and has now decreased to a level almost equal to that of the other national identity groups in For those feeling more Scottish, it seems that the relevance of the general view that wealth is unequally distributed on the decision to support 46

21 Attitudes towards Inequality and Scottish Independence Figure 7 Effect of agreeing that the government should redistribute income from the betteroff to those who are less well off on the probability of supporting independence by national identity Scottish independence is now very low. However, these results should be interpreted with caution as there were no statistically significant differences and therefore may just be random fluctuations over time. Figure 7 shows the results by national identity for the effect of agreeing that the government should redistribute income from the better-off to those who are less well off on the probability of supporting independence. Here we see again that the effect of agreeing that the government should redistribute income on the probability of supporting independence is higher among the groups identifying as more Scottish. During 2010 when the Conservative- Liberal Democrat coalition was formed following the UK general election, it is of interest to note the decline in the effect of agreement about the 47

22 Claire L. Niedzwiedz and Mor Kandlik-Eltanani government s role in income redistribution on support for independence among all national identity groups, but steeper amongst the groups with a more Scottish identity. It is also interesting that there has been an increase in the effect of agreeing the government should redistribute income on support for independence from 2011 to 2012 (after the SNP formed a majority Scottish Government) among the more Scottish than British and especially the Scottish not British groups. Whereas, among the other national identity groups the effect remained reasonably stable. For the two groups identifying as more Scottish, we see the opposite trend to the one described earlier relating to attitudes towards wealth inequality a rise in the relevance of the government s role in income redistribution for the decision to support Scottish independence. Again, these results should be interpreted with caution as no statistically significant interactions were found. Discussion Our examination of attitudes towards income and wealth inequality, income redistribution, and taxation has found that individuals holding more left-wing attitudes towards these issues were more likely to support independence. The strongest associations were observed for the agreement that the government should redistribute income and that wealth is unequally distributed. Individuals agreeing that income inequality is too high within the United Kingdom and that the government should increase tax were more likely to support independence compared to those who disagreed or held a neutral opinion. However, for these variables the associations were weaker and could be explained by party identification. On the one hand, party identification could be interpreted as the cause for both these elements, meaning there is no causal relationship between attitudes towards income inequality and taxation and support for independence, but that it is party identification that primarily influences support for independence. A different interpretation would point out that a choice of party identification is based not only on traditions and social relations, but also on ideological views. This interpretation would suggest that there is a causal relationship, and that attitudes towards social inequality influence the decision to support independence, via political party identification. However, in the absence of longitudinal data, these remain hypotheses. Party identification attenuated, but did not completely explain, the positive associations between the wealth inequality and income redistribution variables, and support for Scottish independence. This suggests that party identification is not the direct cause of 48

23 Attitudes towards Inequality and Scottish Independence these particular relationships as there is an overall association between attitudes towards wealth inequality and income redistribution, and support for independence, which remains unexplained. This of course could be due to factors that we have not included in the statistical models. However, coming back to our original hypothesis, we feel this may be because some individuals, who value a more equal society, feel an independent Scotland may have this potential. This may be because the Scottish Government would be in a better position to be able to alter mechanisms that could have the potential to reduce social inequalities. The relationship between the agreement that wealth is unequally shared and support for independence fluctuated over time. Most notably, there was no difference in the likelihood of supporting independence between those agreeing and disagreeing that wealth is unequally distributed during This has not been observed since 2005 and was apparent across all national identity groups. The reasons for this potential divergence require further exploration. Previously, the effect of agreeing that wealth is unequally shared on support for independence was stronger in the more Scottish groups. The 2013 data will help to elucidate whether this is a continuing trend. In contrast, agreeing that the government should redistribute income continued to have a strong effect on support for independence during The effect was strongest in the two groups who identified as more Scottish, especially the Scottish not British category. Whether this represents an increasing trend of more left-wing nationalism is difficult to tell without the 2013 data and could merely be a random fluctuation over time. Further complicating the picture is the change in the distribution of the Moreno scale in 2012, which could be seen as an effect of the 2012 Olympic Games: the more British categories have gained popularity at the expense of the more Scottish ones. This could have increased the divergence between the groups in 2012, which was seen in Figure 7, because the remaining Scottish groups are smaller, and thus are more extreme compared to the other groups. It will be interesting to see whether there is a similar effect working in the opposite direction during 2014, the year of the Glasgow Commonwealth Games. The differing patterns observed for the questionnaire items about wealth inequality and income redistribution over time require additional investigation. It could be that more general views about social inequality are less important for the decision to support independence. The income redistribution questionnaire item asks specifically about whether the government should intervene to redistribute income from the richer in society to the poorer. Thus, it is perhaps attitudes towards the government s role in reducing income and wealth inequality that are most important for 49

24 Claire L. Niedzwiedz and Mor Kandlik-Eltanani independence support. The nearing 2014 referendum might focus the attention of voters on concrete government actions, rather than on general opinions regarding income and wealth distribution. If this is the case, we could expect this trend to increase in the 2013 wave of the Scottish Social Attitudes survey. However, this does not explain why attitudes towards increasing taxation were less important for the decision to support independence. Perhaps individuals feel they would be more likely to be worse-off economically if taxes were to increase, but think that with income redistribution they may be less likely to lose out, and perhaps even benefit individually as they themselves feel less well-off. The taxation question also asks about whether the government should increase taxes and spend more on areas such as health and education. It could be possible that individuals agree that the government should increase taxes in order to redistribute income or wealth, but disagree that the government should particularly increase social spending. Our study has a number of strengths including the use of representative samples of the Scottish population and the inclusion of important confounding factors in our statistical models. However, it should be noted that the models simplify the complex decision to support independence. We have focused on attitudes towards social inequality, but it is likely that other factors, such as attitudes towards immigration and economic uncertainty also contribute. The 2013 survey data will allow our hypotheses to be explored further, especially in terms of whether the Yes Scotland campaign has been successful in persuading those who value a more equal society that an independent Scotland may have this potential. Nevertheless, the finding that some of our results were not completely explained by party identification has important implications for the framing of the campaigns on both sides of the debate. References Bond, R. (2000). Squaring the circles: Demonstrating and explaining the political nonalignment of Scottish national identitiy. Scottish Affairs, 32, Bond, R., and M. Rosie (2002). National identities in post-devolution Scotland. Scottish Affairs, 40, Bonesmo Fredriksen, K. (2012). Less Income Inequality and More Growth Are they Compatible? Part 6. The Distribution of Wealth, OECD Economics Department Working Papers, No. 929, OECD Publishing. 5k9h28t0bznr-en [Accessed 2 November 2013] Curtice, J. (2011). So Who is Winning the Debate? Constitutional Preferences in Scotland after Four Years of Nationalist Government. Scottish Affairs, 14,

25 Attitudes towards Inequality and Scottish Independence Curtice, J. (2013). Who Supports and Opposes Independence and why? Edinburgh: ScotCen Social Research. Dempsie, P. (2013). A Stark Choice Widening Inequality or a Fairer, more Prosperous Scotland. a_fairer_more_prosperous_scotland. [Accessed 2 November 2013] General Register Office for Scotland. (2011) Census: population report. scotcen3.html [Accessed 2 November 2013] Goldthorpe, J. H. (2010). Analysing social inequality: a critique of two recent contributions from economics and epidemiology. European Sociological Review 26, 6, Lynch, J., G. D. Smith, S. Harper, M. Hillemeier, N. Ross, G. A. Kaplan, and M. Wolfson (2004). Is income inequality a determinant of population health? Part 1. A systematic review. Milbank Quarterly, 82, Mooney, G., and G. Scott (2011). Social justice, social welfare and devolution: Nationalism and social policy making in Scotland. Poverty & Public Policy, 3, Mooney, G., and G. Scott (2012). Social Justice and Social Policy in Scotland. The Policy Press. Moreno, L. (2006). Scotland, Catalonia, Europeanization and the Moreno Question. Scottish Affairs, 54, Paterson, L. (2000). Scottish democracy and Scottish utopias: the first year of the Scottish Parliament. Scottish Affairs, 33, Scott, G., and G. Mooney (2009). Poverty and social justice in the devolved Scotland: Neoliberalism meets social democracy. Social Policy and Society, 3(4), Scott, G., and S. Wright (2012). Devolution, social democratic visions and policy reality in Scotland. Critical Social Policy, 32, Scottish Government. (2013). Fiscal Commission Working Group First Report Macroeconomic Framework pdf. [Accessed 2 November 2013] Wilkinson, R., and K. Pickett (2010). The Spirit Level: Why Equality is Better for Everyone. Penguin Books Limited. Williams, R. (2011). Using the Margins Command to Estimate and Interpret Adjusted Predictions and Marginal Effects. CHI11 Stata Conference: Stata Users Group. 51

26 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Appendix Appendix 1 Percentage for or against independence by attitude towards income and wealth inequality, income redistribution, and taxation (full variables) Ordinary working people do not get their fair share of the nation s wealth Agree strongly Agree Neither Disagree Disagree strongly Against 57% 68% 76% 80% 70% For 43% 32% 24% 20% 30% Total (N) 1,996 5,982 3,051 1, Government should redistribute income from the better-off to those who are less well off Agree strongly Agree Neither Disagree Disagree strongly Against 58% 65% 73% 76% 77% For 42% 35% 27% 24% 23% Total (N) 1,510 3,953 3,386 3, Thinking of income levels generally in Britain today, would you say that the gap between those with high incomes and those with low incomes is... Too large About right Too small Against 69% 76% 59% For 31% 24% 41% Total (N) 8,415 1, Suppose the government had to choose between the three options...which do you think it should choose? Reduce tax Keep same Increase tax Against 61% 72% 69% For 39% 28% 31% Total (N) 520 4,632 6,173 52

27 Attitudes towards Inequality and Scottish Independence Appendix 2 Logistic regression models testing the interaction between survey year and attitude towards wealth inequality and income redistribution Model 1 Agree wealth is unequally distributed Model 2 Agree government should redistribute income Odds ratio [95% CI] Odds ratio [95% CI] Independent variable (main effect) a [0.97,1.80] [0.99,1.84] Year b *** [0.48,1.00] [0.42,0.73] * 0.75* [0.45,0.91] [0.57,0.98] *** 0.45*** [0.27,0.64] [0.33,0.61] *** 0.61*** [0.35,0.72] [0.46,0.81] *** 0.53*** [0.33,0.68] [0.40,0.70] * 0.62** [0.41,0.93] [0.45,0.86] * 0.44*** [0.44,0.96] [0.31,0.62] Interactions [0.53,1.28] [0.72,1.67] [0.86,2.07] [0.70,1.64] [0.84,2.32] [0.93,2.33] 53

28 Claire L. Niedzwiedz and Mor Kandlik-Eltanani Appendix 2 (Cont.) Model 1 Agree wealth is unequally distributed Model 2 Agree government should redistribute income Odds ratio [95% CI] Odds ratio [95% CI] * 1.47 [1.11,2.78] [0.94,2.32] [0.81,2.03] [0.69,1.67] [0.82,2.21] [0.86,2.17] [0.54,1.42] [0.97,2.54] Observations 9,293 9,300 CI = confidence interval; *p < 0.05, **p < 0.01, ***p < a reference category is disagree or neutral in both models; b reference category is year Both models are controlled for age, gender, education level, and national identity. 54