CAIRN-INT.INFO : International Edition

1Thirty years ago, the discovery of the “asset effect” by Jacques Capdevielle and Élisabeth Dupoirier had quite an impact on the field of political science, piquing the interest of numerous researchers, both in France and abroad, in an approach that was considered “heuristic”. [1] In the introduction to his 1981 work France de gauche, vote à droite, which analyzes the left’s defeat in the 1978 legislative election, Alain Lancelot writes the following:


“To my mind, the most original contribution of this book is that it proposes a new explanation by highlighting a new key variable for the study of elections: the possession of assets. It is the type and amount of assets owned – whatever the level of income – that separates supporters and opponents of possible alternatives to the left. Whatever the amount of income, the left always wins out over the right among those who do not have any assets at all, and the right always wins out over the left for those who have two or three types of assets.” [2]

3Inspired by this theoretical advance in France, research on the subject was also carried out in the United Kingdom where the political situation at the time provided an ideal setting for this to happen. Margaret Thatcher, who came to power in 1979, was leading a vigorous campaign aimed at promoting a kind of popular capitalism. She wanted all British citizens to become home-owners and also own shares of some kind. The overarching idea of her policy in this area was to develop a kind of “life ethics” based on the principle that homeowners are more likely to respect the public good and society than non home-owners. In 1998, a conference bringing together top experts in electoral sociology was held at Oxford University. One of the themes raised at the conference touched specifically on the impact of increasing rates of asset ownership on British elections after eight years of Thatcherism. Monica Charlot demonstrated that “this conscious policy of transforming the social environment” had a definite impact on voters ’ choices. During the 1987 legislative election, the diversification of assets played strongly in favor of the conservative vote, to the detriment of the Labour Party.


“The effect of asset ownership on current voting behavior is apparent in Great Britain as it is in France. It has begun to blur the traditional factors that sway the vote significantly, while contributing to a clear differentiation between conservatism and the labour movement.” [3]

5Thus, an absolute majority of workers owning at least three types of assets voted for the Conservative Party, while those who only possessed two types or less clearly preferred the Labour Party. [4]

6Though it was revisited in the 1980s, the asset variable has fallen almost completely into oblivion in the interim. In France, no one since then has undertaken any in-depth research using actual survey data, as evidenced by a reading of the archives of the Revue française de science politique. [5] Although Guy Groux and Catherine Lévy have carried out research on “worker ownership” their approach is more historical and sociopolitical. [6] No researcher has since taken on the task of measuring the effects of an individual’s poverty or relative wealth on their political tendencies and voting choices. Only in the United States has the old tradition of “econometric” analysis lived on, a method that makes use of “macro” logic to attempt to measure the electoral impact of economic variables. The hypothesis that underlies these works is the following: if a country’s economic indices are good, voters will support the existing government; if the indices are bad, they will tend to vote for the opposition. [7]

7There is good reason to return to an analysis based on the relationship between assets and voting behavior in France. Two main hypotheses deserved to be tested, both of which argued for a mitigation of the political impacts of this variable in certain respects. The first approach was to determine whether the asset effect observed during the 1978 legislative election was not in fact linked to the political situation at the time. Indeed, within a left/right binary opposition, and with a Communist Party that still received some 20 % of the vote, the left inspired “great fear” among asset owners. Their fear was that if the left came to power, it would introduce measures which would have a negative impact on the patrimony not only of the “rich” but also of anyone who owned any assets at all. In 1978, the fear of “collectivism” somehow survived the downfall of the programme commun. Today, since the Communist Party has become marginal, and the socialist left is no longer “revolutionary”, it might have been assumed that the strength of the asset effect had decreased in intensity, as asset owners became more willing than before to support candidates from a left-wing party deemed to have become more “administrative” than radical and therefore no longer “frightening”.

8The second hypothesis led to a blurring of the asset effect, which followed from the spread of property ownership among the most diverse sectors of the population, including the working class. In 2007, 57.5 % of households owned their own home, compared to 46.7 % in 1975. [8] Once it became usual, commonplace even, to possess at least two types of assets – in particular owning one’s own home and having a savings account – a dilution of the political impact of asset accumulation seemed inevitable. On the other hand, property and home-ownership were important political issues in the 2007 presidential campaign. Nicolas Sarkozy deliberately focused on the wealthy, as well as anyone – and there were many – who wanted to buy a home, in order to sway them to the right. During the housing Convention which took place on 14 September 2006, he proposed the idea of “property as security” against economic peril in particular, and promised numerous initiatives to promote homeownership. [9] Although the financial crisis had not yet occurred, his aim was to appeal to voters who still shared a somewhat idealized vision of economic liberalism which Sarkozy had taken on board. Thus, in 2007 property ownership served as a strong ideological marker for the right.

9The concept of poverty (or wealth) is nonetheless difficult to define. Indeed, there is no consensus as to its definition, whether among economists, sociologists, or political scientists. [10] Poverty is in effect a “multidimensional phenomenon that does not merely correspond to an indicator of the rate of monetary poverty”. [11] Over the course of this research, wealth was measured using an indicator that combines two elements: income per unit of consumption, and the accumulated assets of the household. This allowed a distinction to be made between “rich” and “poor” in terms of their income on the one hand, and between “rich” and “poor” in terms of their accumulated assets on the other hand. The decision to study income and assets at the level of the household rather than the individual, reflected the work already carried out by economists such as André Babeau and Dominique Strauss-Kahn some time ago, as well as the most recent Insee [12] statistics and the work cited above carried out at Cevipof. [13] Furthermore, the choice of the household as the statistical unit corresponds to more recent social developments: the number of couples in which both partners work fulltime has increased significantly.

10Previous electoral surveys did not include indicators that would enable the adoption of a rigorous approach to household-based standards of living. The 2007 Cevipof post-presidential survey used here filled this methodological gap. The following two questions were also added to the survey: “How many people currently live in your household, counting everyone, including yourself?” and “How many of these are children under the age of fourteen?” Following the example of the Insee, these two indicators, combined with the indicator on income, allow the “standard of living” of responding households by consumption unit to be calculated. [14] It then became possible to determine which households were living below the poverty line. In our survey, this level is fixed at 900 euros per month, a level similar to that fixed by the Insee (908 euros in 2007).

11Does the relative wealth of individuals – understood here as both money and property – bring about certain types of political and voting behavior? This question which is fundamental to our research is all the more pertinent in the context of a long economic crisis that began nearly 40 years ago. As a result of this crisis more economically vulnerable individuals have been pushed below the poverty level. However, others have been encouraged by it to acquire “reserves” in order to cope with the hazards of economic insecurity. The data presented in this article are taken from the 2007 Cevipof post-presidential survey. It is obviously too early to determine the impact of the current financial crisis which started in fall 2008. The focus will therefore be on measuring the political effects of relative wealth in the context of the economic restructuring which began in the mid-1970s as a result of the acceleration of globalized capitalism. Initially, the implementation of redistributive policies limited the social effects of the crisis. However, this is no longer the case. Inequality and poverty have been on the rise since the 1990s. In 2006, the Insee estimated the number of people in France affected by “relative monetary poverty”, to be almost 8 million (13.2 % of the population) compared to 7 million in 2002.

12Concepts already proposed by the Cevipof will be used here: thus, the terms “poor asset owners”, “non-owners”, “relative poverty”, and especially “asset accumulation strategy”, which refers to a system of (right-wing) values will be used. However, it is also true that certain social criteria determine whether an individual is likely to accumulate assets or not, as will be seen below (cf. Table A in the appendix).There are many more opportunities to acquire diversified assets for men, for two-parent households, for graduates and for those with permanent, full-time jobs with a high social standing. On the other hand, although manual workers (22 % of the sample) are over-represented among the non-asset-owning “poor” (29 %) and under-represented among the “rich” with diversified assets (17 %), a portion of them have succeeded in accumulating assets. The ownership of assets is linked to certain types of attitudes and behavior. Developing a strategy for the accumulation of assets results from a double logic which is both social and ideological and it is impossible to know which of these is most important.

Measuring wealth: the income-asset index

13The 2007 post-presidential survey by Cevipof-Ministère de l’Intérieur was used to study the combined effects of income level and asset accumulation strategy on political attitudes and behaviors. The survey supplies the set of indicators needed to build a detailed index. [15]

14It also provides the information on income needed to determine the standard of living of each person interviewed, by consumption units (CU) as per the Insee’s practice. [16] knowledge of each household’s incomes and composition (number and ages of people living in the household) is needed for this process to be completed. [17] The prevailing principle of using units of consumption is that the household’s standard of living should not be considered as a simple linear combination of income and size of the household. On the one hand, living under the same roof allows for economies of scale in household expenses (e.g., rent, telephone service, electricity bill) and on the other hand, the financial costs of raising a child differ according to age. Thus, whereas the first household adult counts for 1 CU, the rest (over age fourteen) count for 0.5 CU, and children under the age of fourteen count for 0.3 CU. The income indicator was dichotomized using an objective measure of wealth as threshold: the 2007 poverty line. This was done to avoid statistical problems with sample size when combining the standard of living with the asset accumulation strategy during the analysis. [18] The first criterion for building the poverty index thus separates the population interviewed into those below and those above the poverty line (34 % and 66 % of the sample, respectively).The Cevipof survey also provides information about all the different types of assets that the household possesses: home, second home, stocks and shares and Type A savings account. [19] The number of assets per household (ranging from zero to four), was tallied and, again for reasons of sample size, people who owned between two and four types were grouped together. [20] Finally, the index used in the survey to gauge the level of relative wealth results from the combination of these two elements, resulting in six distinct modalities (Table 1). This income-asset indicator thus differs from the one designed by the Cevipof for the 1978 survey. By way of reminder, this earlier survey was limited to investment assets only, [21] excluding primary residence. Furthermore, the income categories taken into account were not indexed according to the poverty line. These categorizations are based on objective dichotomy criteria (above and below the poverty line), and on statistical constraints imposed by the sample size. As is often the case, other choices might have been made. However, a more complex breakdown of the income variable by unit of consumption does not change the general conclusions presented in this article. This will become clear in the last section when the logistic regression method is applied.

Table 1

Distribution of the income-asset index (sample size and percentages)[22]

Table 1
No of assets Sample size Percentage Below the poverty line None 227 6 One 575 15 Two to four 501 13 Above the poverty line None 283 7 One 844 22 Two to four 1450 37

Distribution of the income-asset index (sample size and percentages)[22]

Table 2

The number of assets according to income bracket (by consumption unit: CU) (% in rows)

Table 2
No of assets ? 0 1 2 3 4 Income level ? Less than 500 euros (n = 546) 21 48 25 6 0* 501-900 (n = 755) 15 41 33 9 1* 901-1,200 (n = 686) 12 39 36 12 1* 1,201-1,500 (n = 623) 17 37 34 11 1* 1,501-1,800 (n = 704) 9 31 35 19 6 1,801 and above (n = 563) 5 23 35 26 10 Total (n = 3,877) 13 37 33 14 3 *Small number of cases.

The number of assets according to income bracket (by consumption unit: CU) (% in rows)

15The ownership of assets is linked to the level of income available: as Table 2 indicates (the gray cells), the higher the income level, the higher the statistical chances of owning diversified assets. However, the type of assets varies greatly according to income (Table 3): a home, followed by a Type A savings account, are the types of assets most likely to be owned by those who are less well-off financially, [23] whereas assets become diversified as the level of income increases. Owning three or four types of assets (i.e. stocks and a second home) remains the privilege of the richest sectors of the population.

Table 3

Composition of assets according to income bracket (by CU) (% in columns)

Table 3
Income level ? - 500 501 -900 901 -1,200 1,201 -1,500 1,501 -1,800 1801 and above Total No of assets ? None 21 15 12 17 9 5* 13 1 element Home 25 25 25 18 18 12 21 Type A account 22 14 13 17 12 8 14 2 elements Home + stocks 1 * 2 * 3 * 3 * 3 * 9 3 Home + Type A account 20 26 29 26 27 20 25 3 elements Home + 2nd home +Type A 2 * 2 * 3 * 2 * 5 5 * 3 Home + stocks + Type A 4 * 6 9 8 12 18 9 4 elements Home + 2nd hm. +stocks + Type A 0 * 1 * 1 1 * 6 10 3 Other 4 * 7 5 7 9 14 8 * Small number of cases.

Composition of assets according to income bracket (by CU) (% in columns)

16Only 13 % of the sample do not possess any assets at all (Tables 3 and 4), whereas 68 % of those interviewed own their home and 58 % have a Type A account. These high percentages of home-ownership and Type A account explain that while many seniors over the age of 65 [24] live below the poverty line because of a low retirement pension, they also own quite a large amount of assets accumulated gradually over the course of a lifetime. Seniors are therefore less likely than other categories of the population to be characterized as being poor in terms of both money and assets. [25]

Table 4

Composition of assets according to socio-professional category ( % in columns )

Table 4
Profession ? Farmer Owners Exec. / L P Teacher. Mid - LEs routine non -manual Worker Student Non - working Tota l No of assets ? None 8 11 8 9 10 17 15 11 16 13 1 element Home 23 20 12 22 19 24 25 9 26 21 Type A account 2 9 10 13 11 15 15 26 13 14 2 elements Home + stocks 11 9 7 4 3 2 2 2 4 3 Home + Type A account 25 19 24 31 29 24 28 20 23 25 3 elements Home + 2nd home + Type A 1 5 4 6 4 3 2 4 2 3 Home + stocks + Type A 18 11 15 9 13 7 6 11 6 9 4 elements Home + 2nd home + Stocks + Type A 4 6 7 2 4 2 1 5 3 3 Other 8 10 1 3 4 7 6 6 12 7 8 Note: Exec. /LP = Executives, liberal professions; Mid -LE = Mid -level employees.

Composition of assets according to socio-professional category ( % in columns )

17However, the double hardship of living below the poverty line without owning any assets most often affects the working population experiencing employment difficulties according to the following categories (Table A in the appendix): unemployed ( +10 points as compared to the mean), working part-time ( +5 points), or on temporary contract ( +16 points). The following factors also increase the probability of double poverty: not having academic qualifications ( +12 points), being female ( +14 points), or raising children alone ( +10 points). This last factor also impacts women more often than men.

Voter turnout and attitudes to politics

18For a long time now, political science has demonstrated that the level of general interest in politics, voting in elections, following of campaigns, involvement in civil society organizations, etc., varies widely according to socio-cultural level. People with lower levels of education tend to be located lower on the social and professional ladder. They are less likely to be interested in politics or to vote, and more likely to distrust political institutions and politicians. This is all the more true during periods of crisis. There is clear evidence that during such periods, it is the most impoverished, both culturally and socially – those who one might expect to be primarily affected by crisis – who are the ones to lose interest in politics, suggesting that they no longer believe that universal suffrage can improve their situation. [26] What emerges is a paradox in which those who would have the most to gain in making themselves heard by their political leaders pull away from politics, and end up forgotten by the system and public policies. [27] This assessment of the effect of social class on voter turnout and the relationship with the political system was confirmed once again in 2007. However, the income-asset index also shows that the fact of owning diversified assets is a strong factor in political integration, even for those whose income places them below the poverty line, and thus most often at the bottom of the social ladder (Table 5).

Table 5

Attitude to politics according to income and assets ( % per cell )[28]

Table 5
Below the poverty line Abov e the poverty line Total No of assets ? 0 1 2-4 0 1 2-4 Attitude to politics ? Has voted in all elections 47 52 65 47 56 61 57 Belongs to one or several associations 19 22 31 23 30 39 31 Talks about politics with family (often + occasionally) 59 61 71 66 73 78 72 “The result of the presidential election will make things better in France” 43 53 61 49 55 62 57 Is interested in politics (A lot, a little) 49 49 56 58 64 72 62 “Democracy works well” (very well, fairly well) 46 50 62 61 64 73 64 Has followed the campaign more or less every day 56 61 61 60 62 71 63 Note: 47 % of the non-asset-owning “poor” state that they have voted in all elections.

Attitude to politics according to income and assets ( % per cell )[28]

19With regard to electoral turnout, the proportion of interviewees who state that they always vote in elections increases steadily as their ownership of assets becomes more substantial. This result is the same for both the “poor” ( +18 points) and the “rich” ( +14 points). Conversely, those who own no assets are most likely to abstain, whether they are above or below the poverty line. Obviously, the amount of assets held plays a larger role than income level alone when it comes to actually turning out to vote. Similar results can be observed for stated voter turnout in both the first and second rounds of the 2007 presidential election. Whereas in the time of censitary suffrage [29] non-landowners were legally deprived of the right to vote, nowadays it is as if they exclude themselves from political citizenship by choosing to avoid the polls. Jacques Capdevielle and Élisabeth Dupoirier made a similar observation when they wrote: “For all income brackets, the holding of a very diversified set of assets results in a lower level of abstention.” [30] They suggested that this result may be explained by the affluent population’s fear of the consequences of a left-wing victory. Today, many years after the left took power in 1981, the permanent link between voter turnout and ownership of diversified assets suggests that the impulse towards electoral civic-mindedness among the wealthy has little to do with the political situation at the time. It seems that sensitivity to the issues at stake during an election – and the impact these issues might have on their own interests – is always stronger among the wealthy than among those who do not own assets, thus motivating them to vote.

20On the other hand, responses to the other questions in Table 5 show a cumulative effect [31] of income and assets in that the degree of political involvement is much higher among the “rich” who hold highly diversified assets than among the penniless “poor”. Whatever indicator is used (involvement in associations, political conversations, interest in politics, following the campaign regularly, etc.), voters who fall below the poverty line [32] are slightly more involved or interested in politics when they have adopted an asset accumulation strategy: one might conclude that in this sense, assets, whether financial or real-estate, can make up in part for a lack of education. [33] Thus, 71 % of the “poor” who have accumulated several types of assets talk about politics within the family, as opposed to 59 % of the non-assetowning” poor” ( +12 points). Likewise, 61 % of the asset-owning “poor” think that the “presidential election will make things better in France”, as opposed to just 43 % of the nonasset-owning “poor” ( +18 points).

21An explanation is needed for the correlation observed between owning diversified assets and a high degree of political involvement. It is possible that those who adopt an asset accumulation strategy become involved in politics because they assume that the outcome of the election might have a negative or positive effect on their strategy. [34] Or the opposite hypothesis according to which the most politicized people are most likely to accumulate assets may also be valid. Although the underlined correlation negates any reasoning in terms of causality, it is clear that the “poor” who hold diversified assets prove at times to be more politicized than the non-asset-owning “rich”. They are more likely to belong to an association ( +8 points), to talk about politics within the family ( +5), and are more likely to believe that “the presidential election can make things better” ( +12). These themes are in fact directly linked to their individual concerns and investments. With regard to the three other more general indicators of politicization – interest in politics, judgment of the role of democracy, and following a campaign – the asset-owning “poor” are in an equal position with the non-assetowning “rich” (Table 5).

22In brief, the existence of even small amounts of assets among the “poor” seems to act as a shield against exclusion from politics. It seems to motivate people to vote, to talk about politics, to become involved in a campaign, and to see voting in the presidential election as worthwhile. Without any investments to “protect” and cultivate, interest in both public affairs and the electoral process withers away. This same mechanism can also be found among the richest population, for whom the asset accumulation strategy reinforces politicization to an even greater degree. It is clear that the level of monetary wealth and the level of diversified assets combine and reinforce one another, thus facilitating a strong interest in politics.

Mistrust, pessimism, protest

23Some authors have analyzed political trust as being an extension of social trust. This is the case for Robert Putnam, for whom “social capital” is one of the conditions for democracy to function properly. [35] The link between the financial capital and property assets owned by individuals, and their attitudes towards mistrust, protest and pessimism, are equally important as will be demonstrated below.

24The mistrust expressed toward the ability of political parties to govern provides a good example. The level of mistrust reaches its peak among the non-asset-owning “poor” (62%), and falls to 48% among the “poor” who hold diversified assets. It reaches its lowest level (39%) among the “rich” who hold two types of assets or more (i.e. a total difference of 23 points). The asset effect is such that it blurs the line between “rich” and “poor”.

Table 6

Mistrust, pessimism, fear of the future and protest, according to income and assets (% per cell)

Table 6
Below the poverty line Above the poverty line Total No of assets ? 0 1 2-4 0 1 2-4 Mistrust, pessimism ? Trusts neither the right nor the left to govern 62 58 48 54 49 39 48 “ Politicians do not care about what people think” 85 80 72 77 74 65 72 Approves of taking over public buildings in protest 53 48 39 44 44 37 42 Is “afraid when thinking about the future” 73 70 69 62 62 54 62 Note: 62 % of the non-asset kowning “poor” state that they do not have confidence in either the right or the left to govern.

Mistrust, pessimism, fear of the future and protest, according to income and assets (% per cell)

25The two types of assets that make up the index have a simultaneous effect on the other items in Table 6. Thus, the probability of believing that politicians do not care about what people think declines within both income categories of “rich/poor” when respondents in each category own one or more types of assets. Furthermore, the tendency to support protest which is strongest among the non-asset-owning “poor” decreases significantly when the latter own assets. Even for those below the poverty line, the fact of owning anything (be it only a Type A savings account or a small home), seems to lead individuals to be less tolerant of public buildings being occupied in protest. It would appear that respect for private property leads to a similar respect for public property.

26Pessimism about the future is doubly dependent on both income level and the asset accumulation strategy adopted (there is a difference of 19 points between the six modalities of the index). A low income level combined with a total lack of assets makes people particularly concerned about the future. The opposite is also true. During an economic crisis, owning diversified assets seems on the contrary to act as a safety net for an individual and their family when faced with the risk of unemployment or a receding of the welfare state. In this sense, asset ownership increases confidence. However, in this case, the fact itself of owning diversified assets is not enough to compensate for low income: people in this situation are still more concerned when thinking about the future than the “rich” who do not own assets.

Political position, voting preference and party identification

27Athough poverty and sympathy for the left act in concert with each other – and wealth and sympathy for the right go hand-in-hand – this general relationship is blurred by the presence or absence of an asset strategy (Table 7). When they are non-owners, both the “poor” and the “rich” define themselves as left-wing or position themselves to the left on the left/right scale. Conversely, as they accumulate more types of assets, the “poor” locate themselves just slightly less to the right than the “rich”. Asset strategy is thus at the heart of political preference and, as will be shown below, is a major predictor of voting behavior.

28Having an asset accumulation strategy is also an inclusive factor that inspires individuals (even the “poor”), to locate themselves – either to the right or left – on the left/right scale, and declare affiliation with a particular party. Although poverty and refusal to take a political position, or poverty and the refusal to opt for one party over another generally go together, this relationship is challenged when the subject owns two types of assets or more. Thus, those who are “poor” because of their low income, but who nonetheless own diversified assets, are twice as likely to choose a party than the non-asset-owning “poor” (21% as opposed to 10%).

29When the two types of poverty, income and assets, are combined this creates variation in declared political identification, either for the left-wing parties, or for the UMP. The proportion of interviewees who identify with one of the left-wing parties goes from 52% for the nonasset-owning “poor”, to 39% for the asset-owning “poor”, and falls to 34% for the asset-owning “rich” (a total difference of 18 points). It should be noted however, that the connection between poverty and party affiliation is more pronounced with respect to the far left and the Communist Party whereas it is much less obvious with respect to the Socialist Party and its allies. [36] For the right, the relationship is at its strongest with respect to the UMP. The number of people who declare that they identify with the UMP increases in direct proportion to their degree of wealth, both income-wise and asset-wise: this number increases from 9% for the non-asset-owning “poor” to 28% for the asset-owning “poor” and reaches a peak of 37% among the asset-owning “rich” (+28 points). Trust in the right’s ability to govern also varies as a result of the double effect of income and asset-owning. Although trust is weak among the non-asset-owning “poor”, it increases incrementally with the five other terms of the index, going from 18 % up to 42 %( +24 points). However, it should be noted that trust in the left’s ability to govern is far less likely to increase as people’s poverty (income or asset related) increases which, in hindsight, seems to explain Ségolène Royal’s defeat. [37]

Table 7

Political ideology according to income and assets (% in columns) [38][39][40][41]

Table 7
Below the poverty line Above the poverty line Total No of assets ? 0 1 2-4 0 1 2-4 Political ideology Political identification Fairly left 41 37 34 40 39 32 35 Fairly right 17 23 40 32 33 44 36 Neither one nor the other 42 39 26 28 28 24 28 Self-positioning Left 33 33 26 30 32 27 29 Center 13* 13 14 20 17 18 16 Right 17 20 36 26 29 39 31 Neither right nor left 37 33 23 24 22 16 22 Preference for government Trust in the left 19 20 16 18 18 18 18 Trust in the right 18 22 35 28 32 42 33 Neither one nor the other 62 58 23 24 22 16 22 Party identification Far left/Communist Party 21 12 10 10* 9 6 9 PS/MRC/RG/Verts 31 30 29 32 31 28 30 Total left-wing parties 52 42 39 42 40 34 39 UDF 8* 5 9 11 11 12 10 UMP 9* 11 10 10* 6 6 8 FN/MPF/CPNT 21 23 10 13 14 10 14 Accumulation of left-wing characteristics None 39 44 49 41 41 48 45 1 or more 61 56 51 59 59 52 55 Accumulation of right-wing characteristics None 66 61 45 47 50 37 47 1 or more 34 39 55 53 50 63 53 * Small number of cases.

Political ideology according to income and assets (% in columns) [38][39][40][41]

30Finally, when the link is established between an individual’s right-wing tendencies and the asset strategy chosen, it seems clearer than ever that having such a strategy corresponds to a clearly defined value system. [42] The number of those who have one or more right-wing characteristics is low among the non-asset-owning “poor” (34 %), but increases to 55 % among the asset-owning “poor” and reaches more than 60 % among the “rich” with the greatest number of assets (a deviation of 29 points). For left-wing characteristics, the inverse is true though less obvious (there is a deviation of 9 points). There appears to be a strong affinity between asset owners and the right which is translated into votes at election time, while the ideological connection between non-asset owners and the left does not have the same strength. This latter result seems to reflect the fact that the Socialist Party has gradually lost its working-class support, while the Communist Party which used to symbolize the ideological struggle against the affluent has become marginal.

Electoral choices

Table 8

Voting behavior during the first round of the 2007 presidential election, according to income and assets (figures in columns as a % of votes cast )

Table 8
Below the poverty line Above the poverty line Total No of assets ? 0 1 2-4 0 1 2-4 Besancenot/Bové/Buffet/Laguiller/Schivardi 15 11 10 12 10 6 9 Royal 29 30 22 27 29 23 26 Total left (including Voynet*) 44 43 34 39 41 31 37 Bayrou 18 14 17 21 20 21 19 Sarkozy 25 22 34 28 27 37 31 Le Pen/Villiers/Nihous 14 21 16 12 12 12 14 * The left’s total also includes votes for Dominique Voynet, which explains why the sum of “Besancenot/Bové/Buffet/Laguiller/Schivardi” + “Royal” is different from the “Total left”.

Voting behavior during the first round of the 2007 presidential election, according to income and assets (figures in columns as a % of votes cast )

31During the first round of the 2007 presidential election (Table 8), the turnout for the left as a whole reached its highest level among the non-asset-owning “poor” (44 %), falling to 34 % for the asset-owning “poor”, and even farther to 31 % among the asset-owning “rich” (a deviation of 13 points). The vote for the socialist Royal also significantly decreased among each of the six categories in the index, though on a smaller scale (6 points). The income/asset effect on voting behavior can be summarized by remarking that Royal received almost the same support (with a deviation of only two points) from the “poor” as from the “rich” when neither category owned any assets. [43] Likewise, Nicolas Sarkozy received more or less the same support (with a deviation of only 3 points) from the “poor” as from the “rich”, provided that each category owned two or more types of assets. [44] Thus, at the same income level, voter support for Royal decreased as the quantity of respondents ’ assets went up; the opposite effect is true for the Sarkozy vote. This goes to show how much, still today, property assets structure voting behavior and blur income boundaries.

32While poverty and the left-wing vote go hand-in-hand, poverty and radical tendencies also maintain close ties. On the far left, Olivier Besancenot, José Bové, Arlette Laguiller, Gérard Schivardi, and Marie-George Buffet attracted their strongest supporters (15 %) from the nonasset-owning “poor”. This suggests that their “proletarian” status (in the Marxist sense of the term, i.e. those whose only means of subsistence is to sell their labor power) aligns them ideologically with the anti-capitalist cause led by the standard-bearers of Trotskyism and communism. On the far right, Jean-Marie Le Pen, Phillippe de Villiers, and Gérard Nihous find favor among the “poor” with minimal assets (only one element): 21 % of this category voted for one of the three far-right candidates (as opposed to 14 % of the non-asset-owning “poor” and 16 % of the “poor” with diversified assets). This might be classed as a ‘petit blanc ’ [45] vote: those who own little and are financially insecure may well see the far right and notably Jean-Marie Le Pen as the best defenders of their interests, particularly during a time of economic crisis.

33During the second round of the presidential election, the shaping of electoral choices in relation to property assets was even more evident. Given the same income level, fluctuations in the vote according to the categories in the index have a range of close to 20 points for those who live below the poverty line and close to 13 points for everyone else. However, these variations narrow considerably when wealth is considered only in terms of income, controlled for assets (Table 11). [46] Royal obtained significant support (62 %) from the non-asset-owning “poor”. While she still received a majority (53 %) of votes from the “poor” who owned only one type of asset, this figure dropped to a minority figure among the “poor” who owned several elements (43 %). This last group expressed a strong preference for Sarkozy. Much comment has been made on the fact that Royal failed to win over the popular vote. However, she obtained significant support from the most impoverished of the “poor”, those who fall below the poverty line and own no assets. On the other hand, the asset-owning “poor” supported Sarkozy at a rate similar to that of the asset-owning “rich” (57 % and 59 %, respectively), which shows that owning diversified assets trumps any inequality introduced by income.

Table 9

The second round vote in the 2007 presidential election, according to income and assets (figures in columns as a % of votes cast)

Table 9
Below the poverty line Above the poverty line Total No of assets ? 0 (190) 1 (472) 2-4 (438) 0 (238) 1 (711) 2-4 (3,328) (1,279) Royal 62 53 43 54 50 41 47 Sarkozy 38 47 57 46 50 59 53 Note: statistics in parentheses indicate the number of respondents.

The second round vote in the 2007 presidential election, according to income and assets (figures in columns as a % of votes cast)

Table 10

The second round vote in the 2007 presidential election, according to income bracket (by CU) (figures in columns as a % of votes cast)

Table 10
? -500* (448) 501-900 (651) 901-1,200 (579) 1,201-1,500 (528) 1,501-1,800 (614) 1,801 + Total (508) (3,328) Royal 48 53 45 47 46 42 47 Sarkozy 52 47 55 53 54 58 53

The second round vote in the 2007 presidential election, according to income bracket (by CU) (figures in columns as a % of votes cast)

Table 11

The Sarkozy vote in the second round of the 2007 presidential election, according to income and assets (in point differences)

Table 11
Effect of assets, at the same level of income Below the poverty line + 19 points No assets Above the poverty line + 13 points 1 type of asset 2-4 types of assets Effect of income, at the same level of assets + 8 points + 3 points + 2 points

The Sarkozy vote in the second round of the 2007 presidential election, according to income and assets (in point differences)

34Though there is no better predictor of the vote than ownership, the type of asset owned is most definitely not neutral (Table 12). Votes for Sarkozy during the second round of the 2007 presidential election were more likely to be cast by home-owners than by people whose only asset was a Type A savings account (52 % versus 44 %, respectively). Owning only a Type A account, common among the working classes, remains a characteristic of the leftwing vote. Owning one’s own home leads in the opposite direction, toward the right. In retrospect, it is easier to understand why, eight months before the presidential election, Sarkozy organized a convention on housing under the promising title: “Fighting economic insecurity: home ownership for all”. [47] The Sarkozy vote reached 63 % among those who own a home and stocks and even touched the 70 % mark among those who own all four types of assets: a home, second home, stocks and a Type A account.

Table 12

The second round vote in the 2007 presidential election according to number and type of asset (figures in columns as a % of votes cast)

Table 12
None (444) Type A acct. (475) Home (684) Home +Type A (856) Home + stocks (133) Home + 2nd home +Type A (116) Home +stocks +Type A (326) Home + 2nd home +stocks +Type A (120) Total (3,433) Royal 57 56 48 45 37 44 38 31 47 Sarkozy 43 44 52 55 63 56 62 69 53

The second round vote in the 2007 presidential election according to number and type of asset (figures in columns as a % of votes cast)

35This suggests that the best slogan for a right-wing candidate in the next presidential election might be “Home ownership for all”, similar to the “Get rich” slogan that Guizot held so dear. [48] However, it might also be argued that a left-wing candidate who successfully covers the topic of assets might also rally significant support. Until now, the socialist left’s relationship to money and property has not been the focus of much reflection which has only served to encourage those with minimal assets to vote to the right (or the far right), even when their low income places them in the “poor” category. In 2007, they were even more likely to support the UMP because it was embodied by Sarkozy, who clearly expressed his desire to ensure that the money and social success of asset owners be seen in a positive light.

36The explosion of the financial crisis may well inspire the French to pursue even more a diversified asset accumulation strategy as the only way of protecting themselves from unforeseen difficulties such as unemployment or the shrinking of the welfare state. In such a context, the left might well strengthen its connection to the working classes (and the middle classes) by establishing a program including measures to help even the smallest savers to buy their own home. For 30 years now, the theme of property as “financial security” has, without a doubt, been promoted more often by the right than by the left. [49]

Attitudes towards economic issues

37In reality, the differences in voter behavior between the wealthy and the poor reflect the opposition between two paradoxical ideologies. Table 13 clearly shows how both income and asset accumulation strategies shape opinion with respect to economic policy, Europe and globalization. The effects of these two variables are often cumulative. At one of the extreme ends of the left/right scale, the non-asset-owning “poor” are the most favorable toward an economic policy that takes workers ’ interests into account through social measures. They are also the most favorable toward a redistributive fiscal policy. They have reservations about what they perceive to be a neoliberal Europe and globalization leading to negative consequences for their jobs and social benefits. Only a small minority of them perceive the word “profit” positively. At the other end of the scale, the multi-asset-owning “rich” appear to be the strongest advocates of the liberal model. They favor economic competitiveness and free enterprise and are very opposed to the raising of taxes. They see globalization as an opportunity and membership of the European Union as a good thing. Finally, they see the word “profit”, in a positive light, a fetishistic concept that connects them symbolically to the ethos of capitalism. Between these two extremes, a range of opinions according to the different categories of the index can be observed, thus showing the cumulative effect of income and assets on the expression of economic choices. For example, 58 % of the non-asset-owning “poor” perceive globalization as a threat. This figure then decreases over the course of the index’s five remaining categories dipping to 36 % among the multi-assetowning “rich” (-22 points).

Table 13

Attitudes towards economic issues according to income and assets (figures in columns in %)

Table 13
Below the poverty line Above the poverty line Total No of assets ? 0 1 2-4 0 1 2-4 Attitudes towards economic issues ? Minimum wage… Encourages me to do nothing 51 57 64 54 57 57 57 Is a much-needed boost 49 43 36 46 43 43 43 Economic priority/Salaried employees Economic competitiveness 13 18 31 25 27 43 31 Improvement employees’ situation 87 82 69 75 73 57 69 Raise taxes for those who earn more than € 4.000/month Total: agree 62 61 57 55 50 38 49 The state’s role to toward private companies More freedom for companies 36 44 48 46 44 56 49 Tighter control 64 56 52 54 56 44 51 Profit Total: positive 46 50 55 54 55 63 57 Globalization is… An opportunity 15 18 21 19 23 31 24 A threat 58 50 49 44 45 36 44 Membership of the European Union A good thing 33 37 48 46 50 61 51 Fears about the European Union Less social protection 79 75 73 68 70 61 68 More unemployment 84 78 78 71 67 59 68

Attitudes towards economic issues according to income and assets (figures in columns in %)

Does the asset effect transcend “heavy” variables?

38Two distinct methods of analysis were used to obtain a more precise idea of the connection between political behavior and asset strategy. Firstly, a four-way table was designed to test whether the income/asset effect can diminish the effects of significant variables such as membership of a socio-professional group, level of education or religion. Secondly, in order to verify whether the effect observed was maintained once the classic socio-demographic characteristics of the respondents (in particular, age, gender, and level of education) were taken into account, a series of logistic regression analyses was carried out.

39Using a simplified income/asset index (reduced from six to four modalities because of the sample size, to avoid small numbers of cases), the votes cast by routine non-manual employees and workers during the second round of the 2007 presidential election were identified and placed in a separate table (Table 14). The sample size was 865 and 705 people respectively. [50] Using this new index, it becomes clear that the asset effect partially cancels out the effect of the social group to which a voter belongs. A majority of “poor” employees voted for Royal (54%) when they owned just one type of asset, whereas they swung sharply in favor of Sarkozy (6%) when they had accumulated at least two types of assets. This shows completely opposing political views within a single social group. Although a majority of “rich” employees whatever their asset strategy (Table 14) voted for Sarkozy, this vote reached its peak among multi-asset-owning voters (6% and 5%, respectively, a deviation of 12 points). The key variable here is in owning diversified assets, even if income level has its own independent effect on the vote.

40A similar situation can be observed among workers. Even though Sarkozy was able to boast 52 % of the “working-class” vote during the second round of the 2007 presidential election, this average is deceptive. In reality, only a majority of workers whose wealth was made up of two factors – income and assets – voted for him (61%). The other categories of workers – the asset-poor or income-poor – preferred Royal to him (Table 14). Though the socialist left benefited at one time from the gentrification of the working class, [51] this is no longer the case today. Instead, this phenomenon worked to the benefit of the UMP, while the Socialist Party attracted the most impoverished workers. [52] This result highlights the social and political diversity of the working class, which is less homogeneous than ever.

41It is clear that there is nothing automatic about the political orientation of a social group. The “class vote” of routine non-manual employees and workers is easily overcome when voters own diversified assets. Even though the vote remains at least partially determined by one’s position within the production apparatus, it is also determined by the amount of income, and even more so by the types of assets owned. It could be argued that while there is no direct or “necessary” relationship between social class and voter tendency, there is an indirect relationship between the two mediated by asset strategy.

Table 14

The routine non-manual employees’ and workers’ vote during the second round of the 2007 presidential election, according to income and assets (figures in columns as a % of votes cast)

Table 14
Routine non-manual employees Below the poverty line Above the poverty line Total RNM employees (865) No of assets ? 0-1 2-4 0-1 2-4 (187) (135) (289) (254) Royal 54 39 49 37 45 Sarkozy 46 61 51 63 55 Workers Below the poverty line Above the poverty line Total workers (705) No of assets ? 0-1 2-4 0-1 2-4 (173) (104) (216) (212) Royal 53 53 51 39 48 Sarkozy 47 47 49 61 52

The routine non-manual employees’ and workers’ vote during the second round of the 2007 presidential election, according to income and assets (figures in columns as a % of votes cast)

42The effects of educational qualification on voting behavior are similarly disrupted by the type of asset strategy an individual pursues (Table 15). While the electoral “norm” for a majority of those without any qualification was to vote for Sarkozy (57%), this norm was overturned by the non-asset-owning “poor”, a majority of whom voted for Royal (54%). Conversely, the norm for holders of a baccalaureate or university degree, which was to vote for Royal (55%), increased among the non-asset-owning “poor” (64%) and decreased among the “rich” who owned diversified assets (50%).

43Furthermore, while the effect of asset ownership was huge among the “poor” with a qualification lower than the baccalaureate (the Sarkozy vote increased from 46% to 61%), it was more restricted when they held a qualification at least equivalent to a baccalaureate. In the latter case, those below the poverty line were less convinced by Sarkozy. A minority of this population voted for him whatever their asset status was (increasing from 36% to 42%, a deviation of 6 points).

Table 15

The second round vote in the 2007 presidential election, according to level of education, income and assets (figures in columns as a % of votes cast)

Table 15
Qualification lower than the baccalaureate Below the poverty line Above the poverty line Total “less well-educated” (2,254) No of assets ? 0-1 (532) 2-4 (355) 0-1 (587) 2-4 (780) Royal 54 39 46 35 43 Sarkozy 46 61 54 65 57 Baccalaureate or higher Below the poverty line Above the poverty line Total “better-educated” (1,075) No of assets 0-1 (130) 2-4 (83) 0-1 (363) 2-4 (499) Royal 64 58 58 50 55 Sarkozy 36 42 42 50 45

The second round vote in the 2007 presidential election, according to level of education, income and assets (figures in columns as a % of votes cast)

44The income-asset variable far from cancels out the effects of religion on voter tendency [53] since, irrespective of their relative income, a greater proportion of Catholics voted for Sarkozy than those with no declared religion. Nonetheless, the effects of asset accumulation strategy on the electoral choices of Catholics are far from negligible (Table 16). While the political norm for Catholics in 2007 was firm support for Sarkozy (63%), the proportion of the vote for him fell to 53% among the poorest Catholics, and increased to 68% among the most affluent in both income and assets. Asset accumulation makes Catholics lean to the right, whereas left-wing Catholics are more likely to be among the “poor” and not home-owners. This relationship between religion and asset status had already been similarly highlighted in the 1978 survey. It is possible to specify here (which the authors of “The asset effect” could not do because of an insufficient sample size) that, regardless of their distribution according to the four groups of wealth, those without a declared religion always opted strongly in favor of Royal: it appears that those who classified themselves as having “no religion” did not allow their vote to be dictated by their monetary income level, or by the extent of their assets. This appears to be an exception to the rule that the asset-wealthy are always close to the right.

Table 16

The second round vote in the 2007 presidential election, according to religion, income, and assets (figures in columns as a % of votes cast)

Table 16
Catholics (practicing and non-practicing) Below the poverty line Above the poverty line Total “Catholics” (2,155) No of assets ? 0-1 (394) 2-4 (316) 0-1 (552) 2-4 (893) Royal 47 35 38 32 37 Sarkozy 53 65 62 68 63 No declared religion Below the poverty line Above the poverty line Total “No religion” (962) No of assets 0-1 (208) 2-4 (98) 0-1 (329) 2-4 (327) Royal 65 65 70 62 66 Sarkozy 35 35 30 38 34

The second round vote in the 2007 presidential election, according to religion, income, and assets (figures in columns as a % of votes cast)

Testing the asset effect through logistic regression

45The next step was to verify that the major tendencies observed in the tables – underlining the significant effect of wealth by means of a four-way index of both income and assets – are not the result of the “hidden” effect of some other variable. Logistic regression analyses were designed to ascertain this. These analyses checked the effects of gender, age, and level of education on all the variables of political behavior studied – in addition to income and assets (see Tables E through G in the appendix).

46Once traditional socio-demographic characteristics have been accounted for the results confirm the sustained effects of both income and ownership of assets on political and electoral behavior. The odds ratios attached to these two variables are significant and particularly high. Thus, asset ownership and income have been observed – ceteris paribus[54] – to have a significant double effect on all the variables presented here – attitude to politics, pessimism and mistrust of politics, voting behavior and political ideology. There is one exception to this: whether the subject has voted in all possible elections. For this last variable, as mentioned above, only the asset effect seems to hold.

47Because of sample size and small numbers of cases, a four-modality index has been used up to now. However, this type of index is not very nuanced, in particular with respect to income by (dichotomized) unit of consumption but also to asset structure. Since logistic regression analysis allows for this, it therefore seems appropriate to rearrange the standard of living into sextiles. [55] The standard of living firstly underwent a logarithmic transformation. This method has the advantage of attributing more weight to small income values, thus giving voice to those who live below the poverty line and who are far less represented in the sample. [56] Likewise, it was possible to divide the number of asset variables into four distinct categories (instead of three as in previous sections). The result is significant and highlights, yet again, the strong stability of these effects. This confirms the value of the income and asset variables in explaining political behavior. [57]

48The analysis will be centered on the Sarkozy vote during the second round of the 2007 presidential election (Table 17). It should be stressed that the tendencies observed here are often identical for the other independent variables considered thus far.

49The initial model (Model 1) takes account not only of income and assets, but also of gender, age, and level of education. It shows the large role played by the level of relative wealth in the Sarkozy vote during the second round of the presidential election.

50Thus, individuals who do not own any assets are twice as unlikely (2.1 times precisely) to have voted for Sarkozy than individuals who own three or four types of asset. Likewise, belonging to the lowest of the six income brackets cuts the chance of voting for Sarkozy by more than 1.5 times (specifically 1.6 times) compared to someone at the top of the income scale (the sixth bracket). Consequently, even though the level of education seems decisive (the odds ratios are high in Model 1), assets and income must also be considered to understand the outcome of the 2007 presidential election.

51In the section above, the importance of religion (a so-called “heavy” variable) in making an electoral choice was stressed. Furthermore, the huge effect of employment status on voting behavior is already well-known. [58] The integration of these two variables into the model – Model 1 – therefore seems essential. This is the reason for the second and third models (to which we have added status – Model 2 – and religion – Model 3).

52An initial observation is that these two variables (religion in particular) make a substantial contribution to explaining the Sarkozy vote: the Cox and Snell R [59] (which indicate the quality of the model), with an initial value of 4.9 %, increases to 6.1 % in the second model, then doubles with the introduction of the religion variable, reaching 12.4 % in the third model.

Table 17

Three logistic regression models of the Sarkozy vote in the second round of the 2007 presidential election (the coefficients indicate the exponential function (exp(b)), representing the odds ratios)[60][61]

Table 17
Model 1 Model 2 Model 3 Cox and Snell R2 0.049 0.061 0.124 Gender Male 0.946 0.863* 0.961 Female Ref. Ref. Ref. Age 18-24 0.430*** 0.613** 1.037 25-34 0.763* 0.943 1.381** 35-49 0.662*** 0.761** 1.064 50-64 0.605*** 0.650*** 0.764** 65 and older Ref. Ref. Ref. Education No qualification 1.950*** 2.391*** 2.474*** BEPC 1.603*** 1.952*** 1.918*** Baccalaureate 1.302** 1.377** 1.360** Higher than the Bac. Ref. Ref. Ref. No of assets 0 0.470*** 0.546*** 0.713** 1 0.586*** 0.635*** 0.727** 2 0.784** 0.861 0.913 3-4 Ref. Ref. Ref. Income by unit of consumption (logarithmic transformation) Q1 – Low standard of living 0.637*** 0.540*** 0.504***(n) Q2 0.657*** 0.568*** 0.504*** Q3 0.827 0.721 0.628 Q4 0.864 0.807 0.726** Q5 0.739** 0.718** 0.667*** Q6 – High standard of living Ref. Ref. Ref. Status Self-employed, entrepreneur 2.571*** 2.471*** Private sector employee 1.352*** 1.321*** Public sector employee Ref. Ref. Religion Practicing Catholic 6.705*** Non-practicing Catholic 2.912*** Other religions 0.957 No religion Ref. Probability that the coefficients are significant: *** p 0.01, ** p 0.05, * p 0.1. Note (n): In model 3, the likelihood of voting for Sarkozy is cut almost in half (1/0.504 = 1.98) among those who have a low standard of living rather than a high standard of living (Ref.).

Three logistic regression models of the Sarkozy vote in the second round of the 2007 presidential election (the coefficients indicate the exponential function (exp(b)), representing the odds ratios)[60][61]

53Furthermore, the reading of the odds ratios in this last model is instructive on several levels. First, religion is a significant fundamental indicator of voting behavior: all else holding constant a practicing Catholic is therefore seven times more likely to have voted for Sarkozy than an individual with no religion. Both employment status and level of education account for lesser differences (odds ratio of about 2.4).

54The religious effect does not cancel out the effects of income and assets, which remain at altogether significant levels. While the asset effect declines gradually between the first and third models (the odds ratios and levels of significance go down), the standard of living is not much affected by the complexity of the first model, and even sees its odds ratio and “significance” increase. The only conclusion is that individual wealth, measured by asset diversity and income level, merit a great deal of attention from anyone who wishes to fully understand how electoral choices are made.


56* *

57Over the course of this study it has become clear that the presence or absence of an asset accumulation strategy is a deciding factor in the attitude to politics and to the vote. The fact of owning diversified assets or not can even blur the distinction between rich and poor based solely on income, and can erase political norms linked to social, cultural, and professional affinities.

58Being poor in assets is linked to a diminished interest in politics, a refusal to locate oneself on the political scale, a lack of party identification and a left-wing vote. It seems that left-wing parties are the only ones to respond to the concerns of those who have only a small income to combat the hazards of social, economic, and medical insecurity during a time of crisis. These voters invest their hopes in the state continuing to support national solidarity and the fight for social equality under a left-wing government. Nonetheless, a significant portion of them cast a protest vote either for the far left or the far right.

59When the income-poor own two types of assets (including housing in particular), they set up a kind of private solidarity based on individual property and their political leanings change completely. They become politically involved, trusting governmental institutions and parties, and have ideological referents that lead them to favor right-wing candidates whom they perceive as providing the best economic security for home-owners.

60The conclusions drawn here are therefore quite similar to those drawn by Cevipof researchers from the 1978 survey, even though the ideological and political context has changed significantly over the last 30 years. Added to this, looser criteria in the definition of ownership have been adopted, to include both investment properties and the primary residence.

61It seems that in the post-industrial societies of the twenty-first century, the political struggle is still rooted in property ownership. This observation contradicts certain analyses, among them the study by the sociologist Ralf Dahrendorf. In Class and Class Conflict, [62] written in the late 1950s, Dahrendorf tried to demonstrate that the struggle is rooted in the unequal distribution of the roles of command and subordination. However, in 2007, owning diversified assets has proven to be the best predictor of the right-wing vote in France, and more precisely of votes for the UMP candidate, Nicolas Sarkozy. Conversely, poverty as defined by lack of assets corresponded to a vote in favor of the socialist candidate, Ségolène Royal. However, these apparent links between assets and politics must not be interpreted in terms of causality. To do so would lead to the conclusion – which it is not our intention to draw here – that politics might be in a dependent or subordinate relationship with economics. [63]

Appendices: complementary tables
Table A

Sociology ofpovertyandwealthaccording to the income-asset index (%)[64]

Table A
Below the poverty line Above the poverty line Total No of assets ? 0 1 2-4 0 1 2-4 Sociology ? Gender Male 34 33 42 53 52 55 48 Female 66 67 58 47 48 45 52 Age 18-24 15 16 13 12 7 10 11 65 and over 14 20 34 10 14 22 20 Familial status Couple without children 15 19 26 19 26 33 27 Couple with children 35 35 34 33 40 36 36 Single without children 24 20 17 34 24 17 21 Single with children 26 25 23 14 10 13 16 Education No qualification 45 49 45 26 25 26 33 BEPC/CAP/CEP 34 34 35 34 38 35 35 Baccalaureate/Bac+2 16 14 17 26 27 26 23 B.A. or higher 5* 4* 2* 14 10 14 9 Socio-professional category Farmer 4* 4* 11 1* 1* 2 3 Business owner 6* 5 6 4* 5 7 6 Liberal professions/Executive 2* 2* 2* 10 10 16 9 Mid-level employee 4* 7 7 20 21 25 18 Routine non-manual employee 32 29 30 35 30 20 27 Worker 29 26 25 23 24 17 22 Other/non-working 24 26 19 7* 9 13 15 Type of Agglomeration Rural 18 27 34 14 26 25 26 Less than 20,000 16 19 19 19 16 17 17 20,000-100,000 17 14 13 14 14 13 14 More than 100,000 34 33 24 35 28 29 30 Paris 16 6 9 18 16 16 14 Type of work contract Permanent 65 70 74 79 85 84 81 Temporary 35 30 26 21 15 16 19 Status 1 Full time 29 24 27 66 56 46 43 Part time 15 13 9 8* 10 7 10 Unemployed 14 8 4* 5* 4 2* 4 Retired 14 24 36 13 18 29 25 Housewife 14 10 8 4* 4 5 6 Student 7* 13 11 4* 4 7 8 Status 2 Self-employed/Entrepreneur 17 14 23 6* 6 13 12 Private sector employee 46 50 49 57 56 54 53 Public sector employee 36 35 28 37 37 33 34 Worker attributes None 32 38 43 40 41 50 43 1 40 35 34 34 32 31 2-3 28 27 24 26 27 19 24 Ability to manage on household income Total: difficult 83 78 65 67 53 33 54 Total: easy 17 22 35 33 47 67 46 * Insufficient sample size.

Sociology ofpovertyandwealthaccording to the income-asset index (%)[64]

Table B

Make-up of assets according to the income-asset index (%)

Table B
Below the poverty line Above the poverty line Total No of assets ? 1 2-4 1 2-4 Type of asset ? 1 element Home 57 56 13 Type A account 40 38 21 2 elements Home + Type A 62 46 25 Home + stocks 5 * 8 3 3 elements Home + Type A + second home 6 * 7 3 Home + Type A + stocks 13 20 9 4 elements Home + Type A + second home + stocks 2* 8 3 Other combinations (rarer) 3 13 5 12 8 * Insufficient sample size.

Make-up of assets according to the income-asset index (%)

Table C

Interest in politics, according to education, income, and assets (figures in columns as a % of votes cast)

Table C
Qualification lower than the baccalaureate Below the poverty line Above the poverty line Total “less-educated” (2,640) No of assets ? 0-1 (655) 2-4 (401) 0-1 (701) 2-4 (883) A lot + fairly 46 52 56 68 57 Little + not at all 54 58 44 32 43 Baccalaureate or higher Below the poverty line Above the poverty line Total “more-educated” (1,239) No of assets ? 0-1 (147) 2-4 (99) 0-1 (426) 2-4 (567) A lot + fairly 64 71 74 77 74 Little + not at all 36 29 26 23 26

Interest in politics, according to education, income, and assets (figures in columns as a % of votes cast)

Table D

Has voted in all elections, according to level of education, income, and assets (figures in columns as a % of votes cast)

Table D
Qualification lower than the baccalaureate Below the poverty line Above the poverty line Total “less-educated” (2,642) No of assets 0-1 (655) 2-4 (402) 0-1 (702) 2-4 (883) Has voted in all elections 49 64 53 60 56 Baccalaureate or higher Below the poverty line Above the poverty line Total “more-educated” (1,241) No of assets 0-1 (148) 2-4 (99) 0-1 (426) 2-4 (568) Has voted in all elections 57 66 53 63 59

Has voted in all elections, according to level of education, income, and assets (figures in columns as a % of votes cast)

Table E

Logistic regression of variables in relation to politics (exp(B)), *** 1%, ** 5%, * 10%)[65]

Table E
Exp(B) Has voted in all elections Belongs to 1 or more associations Talks politics at home (often, sometimes) “The result of the presidential election will make things better in France” Interest in politics (a lot, fairly) “Democracy works well” (Very well, fairly well) Followed the campaign everyday or thereabouts Cox and Snell R2 0.049 0.045 0.047 0.038 0.083 0.055 0.060 Gender Male 1.046 1.405*** 0.960 0.863** 1.654*** 1.369*** 1.348*** Female Ref. Ref. Ref. Ref. Ref. Ref. Ref. Age 18-24 1.182 0.456*** 1.171 0.586*** 0.901 0.695** 0.552*** 25-34 0.387*** 0.446*** 0.936 0.874 0.681 0.584*** 0.507*** 35-49 0.434*** 0.621*** 0.902 0.783** 0.640*** 0.738*** 0.473*** 50-64 0.549*** 0.829* 0.989 0.706*** 0.925 0.874 0.671*** 65 and up Ref. Ref. Ref. Ref. Ref. Ref. Ref. Education No qualifications 0.827 0.402*** 0.412*** 1.897*** 0.341*** 0.485*** 0.417*** BEPC 0.761*** 0.506*** 0.523*** 1.740*** 0.458*** 0.601*** 0.504*** Bac 0.970 0.667*** 0.685*** 1.190 0.631*** 0.831 0.629*** > Bac Ref. Ref. Ref. Ref. Ref. Ref. Ref. No of assets 0 0.580 *** 0.506 *** 0.543 *** 0.455 *** 0.760 ** 0.546 *** 0.756 ** 1 0.760 *** 0.656 *** 0.661 *** 0.619 *** 0.832 * 0.615 *** 0.641 *** 2 0.937 0.846 * 0.838 0.749 *** 0.973 0.862 0.848 3-4 Ref. Ref. Ref. Ref. Ref. Ref. Ref. Income by unit of consumption Q1 1.056 0.904 0.455*** 0.716*** 0.377*** 0.531*** 0.477*** Q2 0.831 0.865 0.522*** 0.682*** 0.454*** 0.578*** 0.629*** Q3 1.022 0.986 0.596*** 0.755** 0.452*** 0.670*** 0.630*** Q4 0.977 1.021 0.544*** 0.880 0.506*** 0.862 0.659*** Q5 0.912 1.117 0.741** 0.889 0.681*** 0.897 0.797* Q6 Ref. Ref. Ref. Ref. Ref. Ref. Ref. Income *Assets interaction Insignificant Insignificant Insignificant Insignificant Insignificant Insignificant Insignificant

Logistic regression of variables in relation to politics (exp(B)), *** 1%, ** 5%, * 10%)[65]

Table F

Logistic regression for the variables of mistrust, pessimism, fear for the future and protest (Exp(B)), *** 1%, ** 5%, * 10%)

Table F
Does not trust the left or the right to govern “ Politicians do not care about what people think” Approves of occupying government buildings in protest Is “afraid when thinking about the future” Cox and Snell R2 0.051 0.055 0.057 0.069 Gender Male 0.840 ** 0.840 *** 1.053 0.500 *** Female Ref. Ref. Ref. Ref. Age 18-24 1.721 *** 1.098 4.451 *** 1.851 *** 25-34 2.015 *** 1.732 *** 3.082 *** 1.907 *** 35-49 1.863 *** 1.672 *** 2.913 1.532 *** 50-64 1.341 *** 1.629 *** 1.904 *** 1.344 *** 65 and up Ref. Ref. Ref. Ref. Education No qualification 1.106 2.018 *** 0.998 1.678 *** BEPC 1.199 * 2.001 *** 0.923 1.594 *** 1.343 Bac 1.083 1.714 *** 0.901 ** >Bac Ref. Ref. Ref. Ref. Noof assets 0 1.760 *** 1.827 *** 1.568 *** 1.294 ** 1 1.489 *** 1.436 *** 1.571 *** 1.220 * 2 1.261 ** 1.063 1.394 *** 1.197 * 3-4 Ref. Ref. Ref. Ref. Income by unit of consumption Q1 2.346 *** 1.930 *** 1.253 * 2.158 *** Q2 2.168 *** 1.888 *** 1.500 *** 2.206 *** Q3 1.974 *** 2.113 *** 1.142 1.775 *** Q4 1.666 *** 1.850 *** 1.218 * 1.774 *** Q5 1.554 *** 1.387 *** 1.177 1.459 *** Q6 Ref. Ref. Ref. Ref. Income * Assets interaction Insignificant Insignificant Insignificant Insignificant

Logistic regression for the variables of mistrust, pessimism, fear for the future and protest (Exp(B)), *** 1%, ** 5%, * 10%)

Table G

Logistic regressions for the variables of political positioning to the right (Exp(B)), *** 1%, ** 5%, * 10%)

Table G
Political affiliation: the right Self-positioning on the right Preferred government: right-wing Right-wing party identification (UDF + UMP) Right-wing attributes: 1 or more Cox and Snell R2 0.051 0.047 0.054 0.064 0.046 Gender Male 1.088 1.061 1.069 1.111 1.028 Female Ref. Ref. Ref. Ref. Age 18-24 0.432*** 0.483*** 0.476*** 0.440*** 0.560** 25-34 0.693*** 0.754*** 0.604*** 0.602*** 0.880 35-49 0.578*** 0.587*** 0.593*** 0.547*** 0.727*** 50-64 0.638*** 0.711*** 0.690*** 0.618*** 0.833* 65 and up Ref. Ref. Ref. Ref. Ref. Education No qualification 1.221 1.285* 1.241 0.703*** 0.831 BEPC 1.195 1.283** 1.183 0.836* 0.899 Bac 1.055 1.099 1.023 0.847 0.755** >Bac Ref. Ref. Ref. Ref. No of assets 0 0.425*** 0.403*** 0.420*** 0.469*** 0.497*** 1 0.489*** 0.465*** 0.522*** 0.557*** 0.498*** 2 0.749*** 0.702*** 0.682*** 0.853 0.754*** 3-4 Ref. Ref. Ref. Ref. Income by unit of consumption Q1 0.522*** 0.535*** 0.410*** 0.408*** 0.519*** Q2 0.625*** 0.591*** 0.496*** 0.486*** 0.603*** Q3 0.746** 0.637*** 0.570*** 0.707*** 0.808 Q4 0.823 0.782** 0.706*** 0.746*** 0.725*** Q5 0.808* 0.729*** 0.678*** 0.796** 0.801* Q6 Ref. Ref. Ref. Ref. Ref. Income *Assets interaction Insignificant Insignificant Insignificant Insignificant Insignificant

Logistic regressions for the variables of political positioning to the right (Exp(B)), *** 1%, ** 5%, * 10%)


  • [1]
    Jacques Capdevielle and Élisabeth Dupoirier, “L’effet patrimoine”, Jacques Capdevielle et al. (eds), France de gauche, vote à droite (Paris: Presses de Sciences Po, 1981), 171-355. See also Jacques Capdevielle, Le fétichisme du patrimoine: essai sur un fondement de la classe moyenne, (Paris: Presses de Sciences Po, 1986).
  • [2]
    Alain Lancelot, “Préface”, in Jacques Capdevielle et al. (eds), France de gauche, vote à droite, 11-12.
  • [3]
    Monica Charlot, “Un effet patrimoine?”, in Monica Charlot (ed.) L’effet Thatcher (Paris: Economica, 1989), 33-51 (49). See also David E. Butler and Dennis Kavanagh, The British General Election of 1983 (London: Macmillan, 1985): the authors show that workers who own their own homes are more inclined to vote conservative than those living in public housing (296-7).
  • [4]
    The types of wealth that Monica Charlot’s study takes into account are: accounts with National Savings and building societies, shares, and unit trusts.
  • [5]
    Note, however, the study of the asset variable carried out by Daniel Boy and Nonna Mayer, “Que reste-t-il des variables lourdes?”, in Daniel Boy and Nonna Mayer (eds), L’électeur a ses raisons (Paris: Presses de Sciences Po, 1997), 118-23.
  • [6]
    Cf. Guy Groux and Catherine Lévy, La possession ouvrière: du taudis à la propriété (19e-20e siècle) (Paris: Éd. de l’Atelier, 1993). See also Hélène Michel’s more general work, La Cause des propriétaires: état et propriété en France, fin 19e-20e siècle (Paris: Belin, 2006).
  • [7]
    Cf. the following among other recent studies: Raymond Duch and Randy Stevenson, The Economic Vote: How Political and Economic Institutions Condition Election Results (Cambridge: Cambridge UP, 2008); Michael S. Lewis-Beck and Mary Stegmaier, “Economic Models of the Vote”, in Russel Dalton and Hans-Dieter Klingemann (eds), Oxford Handbook of Political Behavior (Oxford: Oxford UP, 2007), 518-37: Wouter Van der Brug, Cees Van der Eijk, and Mark N. Franklin, The Economy and the Vote: Economic Conditions and Elections in Fifteen Countries (New York: Cambridge UP, 2007). For France, see: Michael S. Lewis-Beck, “Le vote du ‘porte-monnaie ’ en question”, in L’électeur a ses raisons, 239-61; Jean-Dominique Lafay, “L’analyse économique de la politique: raisons d’être, vrais problèmes et fausses critiques”, Revue française de sociologie, 38(2), 1997, 229-43, and “Analyse économique d’une présidentielle”, Sociétal, 36, Apr-Jun 2002, 4-8.
  • [8]
    Insee, Les tableaux de l’économie française,
  • [9]
    In his closing speech, Nicolas Sarkozy said: “Property provides security against unemployment or job change, it guarantees the standard of living upon retirement, or when retraining for a new vocation and it provides capital that can be passed on to one’s children” (Le Monde, 10 September 2009).
  • [10]
    Cf. Le rapport de l’Observatoire national de la pauvreté et de l’exclusion sociale 2007-2008 (Paris: La Documentation française, 2008), and the work of Serge Paugam, especially La disqualification sociale: essai sur la nouvelle pauvreté (Paris: PUF, 1994).
  • [11]
    Assemblée nationale, “Rapport d’activité au nom de la Délégation aux droits des femmes et à l’égalité des chances entre les hommes et les femmes”, no. 3670, October 2005-February 2007, 7.
  • [12]
    Translator’s note: Insee stands for “Institut national de la statistique et des études économiques” (National Institute of Statistics and Economic Studies).
  • [13]
    Cf. André Babeau and Dominique Strauss-Kahn, La richesse des Français: épargne, plus-value, héritage: enquête sur la fortune des Français (Paris: PUF, 1977); Luc Goutard and Jérôme Pujol, “Les niveaux de vie en 2006”, Insee Première, 1203, July 2008.
  • [14]
    The standard of living is defined here according to standardized international methods (known as the OECD scale). See the first section of this article for details. In the 1978 survey, one of the indicators – composition by age – was not used. However, the authors managed to sidestep this difficulty: since they knew the size of the household, “after the first three people in the household, we applied a single coefficient of. 5 units for each additional person” (In J. Capdevielle et al., France de gauche, vote à droite, 176).
  • [15]
    The sample contains 4,006 people who are representative of the French population registered to vote, aged eighteen and older. The representativeness of the sample is guaranteed by the quota method (gender, age, head of household’s professional activity or former professional activity, level of education) after categorisation by admininstrative region and type of town. The weighted file (socio-demographic and political weighting – first and second rounds of the 2007 presidential election) has been used for the purposes of this article.
  • [16]
    “The standard of living corresponds to the income available to the household (total income of all salaried workers, other income from work, other benefits, minimum state benefits, income from property, net direct taxes) divided by the number of units of consumption (CU). […] Units of consumption are usually calculated according to a modified version of the equivalence scale defined by the Organisation for Economic Cooperation and Development (OECD), which allows 1 CU for the first adult in the household, 0.5 CU for each person aged fourteen and over, and 0.3 CU for children less than fourteen years of age.” Yves Jauneau, “Les employés et ouvriers non qualifiés: Un niveau de vie inférieur d’un quart à la moyenne des salariés”, Insee Première, 1250, July 2009, 4.
  • [17]
    The questions asked in the Cevipof survey are the following: 1) “What is the total monthly income for your household, including salaries, government benefits, and other income?”; 2) “How many people currently live in your home, counting everyone, including yourself?”; 3) “How many of these are children under the age of fourteen?”
  • [18]
    The poverty line set by the Insee in 2007 was 908 euros. However, we used the figure of 900 euros for purposes of legibility. The value of the poverty line corresponds to 60% of the median standard of living (the criterion used by Eurostat and European countries).
  • [19]
    Translator’s note: A livret A (Type A savings account) is tax-free up to a certain amount.The question asked during the survey was: “Do you yourself, or does any member of your household own: (Yes/No) your home / a second home / securities, stocks (shares, mutual funds) / a type A savings account (savings bank, post office)?”
  • [20]
    Very few people living below the poverty line possess two or more types of wealth.
  • [21]
    The types of wealth used by Jacques Capdevielle and Élisabeth Dupoirier are: Type A savings account (tax free up to a certain amount), standard savings account, savings bonds, treasury bonds, corporate bonds, term deposits, home purchase savings accounts, securities (stocks, bonds, unit trusts), and investment properties.
  • [22]
    It is possible for the percentages not to add up to 100 in the tables presented here. This is most often due to rounding, but may also result from the absence of rare categories, such as “don’t know”, “other”, etc. in the tables. The source of all tables is the 2007 Cevipof-Ministère de l’Intérieur post-presidential survey.
  • [23]
    In 2004, 82.6% of households had tax-free savings. Insee, Les revenus et le patrimoine des ménages, Paris, 2009 edition, 137.
  • [24]
    A detailed table of sociological characteristics is available in the Appendices (Table A).
  • [25]
    According to the Insee, 69.5% of French householders between the ages of 60 and 69 own their primary place of residence, compared with 13.7% of those under the age of 30 (Insee, Les revenus et le patrimoine…, 137).
  • [26]
    Thus, according to the 2007 Cevipof-Ministère de l’Intérieur post-presidential survey, only 52% of workers and 55% of routine non-manual employees are interested “a lot” or “a little” in politics, as opposed to 85% of those in liberal professions or executives. This is also the case for 54% of those with no qualification and those who went no further than primary school, as opposed to 86% of those with a university degree.
  • [27]
    Cf. Lo?c Blondiaux, “Les invisibles de la représentation”, Le Monde, 29 April 2009, and his work, Le nouvel esprit de la démocratie: actualité de la démocratie participative (Paris: Seuil, 2008).
  • [28]
    This table and all following tables have been subjected to verification by logistic regression (independent variables: age, gender, education, dichotomized income, wealth in three categories): the results show that, all other things being equal, the effects of income and patrimony are significant.
  • [29]
    Translator’s note: Censitary suffrage refers to the practice of reserving voting rights and eligibility to those who paid the cens, i.e. a certain amount of tax. The practice was abolished in France in 1848, when universal male suffrage was introduced.
  • [30]
    Jacques Capdevielle et al., France de gauche…, 183.
  • [31]
    A “cumulative effect” occurs when twofold poverty (lack of income and of wealth) lead to simultaneous effects (the two elements combine and act simultaneously). The term “single effect” is used when one of the two effects occurs without the other.
  • [32]
    In the sense that the asset-owning “poor” are located at the same level or higher than the non-asset-owning “rich” (though they do not reach the score of the asset-owning “rich”); on the other hand, when the amount of assets owned is equal, income continues to have an effect.
  • [33]
    We have already highlighted that a double handicap (lacking both income and assets) is characteristic of people with no qualifications.
  • [34]
    Certain elements dealt with further on in this article allow this hypothesis to be formulated.
  • [35]
    Cf. Robert Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon & Schuster, 2000); James S. Coleman, Foundations of Social Theory (Cambridge: The Belknap Press of Harvard UP, 1990); Nonna Mayer, “Les consequences politiques du ‘capital social ’: le cas français”, Revue internationale de politique comparée, 10(3), 2003, 381-95.
  • [36]
    Jacques Capdevielle and Élisabeth Dupoirier had already observed, 30 years ago, that the socialist electorate was much less dependent on the combined effects of income and wealth than the communist electorate (in J. Capdevielle et al., France de gauche…, 183).
  • [37]
    Translator’s note: Ségolène Royal was Nicolas Sarkozy’s socialist opponent in the 2007 presidential election.
  • [38]
    Translator’s note: PS = Parti socialiste (Socialist Party); MRC = Mouvement républicain et citoyen (Republican and Citizen Movement); RG = Parti radical de gauche (Radical Left Party); Verts = The Green Party.
  • [39]
    Translator’s note: UDF = Union pour la démocratie française (Union for French Democracy), a centrist party.
  • [40]
    Translator’s note: UMP = Union pour un mouvement populaire (Union for a Popular Movement), politically right.
  • [41]
    Translator’s note: FN = Front national (National Front); MPF = Mouvement pour la France (Movement for France); CPNT = Chasse, Pêche, Nature et Traditions (Hunting, Fishing, Nature and Traditions); all far right parties.
  • [42]
    Design principle for right-wing characteristics: Since we know the political orientation of the respondent, as well as of his/her parents, and his/her domestic partner ( “on the left”, “on the right”, or “neither on the left nor on the right”), we tallied how many of the four people the respondent describes as being to the right. The design principle for the accumulation of left-wing characteristics is identical, tallying this time how many of them are positioned on the left.
  • [43]
    The observation remains the same no matter what level of wealth is considered: -1 point for a single type of asset between the “poor” and the “rich”, and +1 point for two types of assets or more.
  • [44]
    See the above remark.
  • [45]
    Petits blancs” is an expression in French used to refer to poor segregationist whites in the southern US in the 1950s. Those who lived near black ghettos felt their possessions and very existence were threatened by the Afro-American population.
  • [46]
    This is no doubt due in part to the effect of a dichotomized recoding of income. However, it is worth noting that the double prism (income + assets) refines the analysis of the vote considerably. Furthermore, as will become clear later through the use of logistic regression a more detailed income index leads to the same conclusion.
  • [47]
    During the convention held on 14 September 2006 (mentioned above), Sarkozy promised to create a “0% home loan” and to sell off “1% of subsidized housing each year”. He also called for the “reintroduction of a tax credit for mortgage interest”. In this sense, Sarkozy can be seen as the worthy successor of Valéry Giscard d’Estaing who, in his speech at Verdun-sur-Doubs (27 January 1978), declared his wish to “make the French the owners of France”.
  • [48]
    Translator’s note: François Guizot was an important government figure throughout the 1830s and 1840s, becoming prime minister in 1847. His famous advice to the poor was that they should strive to join the middle class and become owners so that they would be entitled to vote.
  • [49]
    The relationship between property and (financial) security/freedom was amply developed by Valéry Giscard d’Estaing: “Additional [financial] security leads people to feel that they have a personal patrimony. The freedom to wait, to choose, to decide, is reinforced by the possession of a ‘reserve ’ that protects against external instabilities.” Démocratie française (Paris: Fayard, 1976), 122. However, the glorification of property as financial security was also promoted (as noted above) by Sarkozy during the 2007 presidential campaign.
  • [50]
    The sample size of the other socio-professional categories is insufficient.
  • [51]
    Cf. G. Groux, C. Lévy, La possession ouvrière…, 191-3.
  • [52]
    It can be assumed that this category partly intersects with routine non-manual employees and unskilled workers, whom we know to be the most prone to poverty (cf. Y. Jauneau, cited above). On the voting tendency of the working classes, cf. Guy Michelat, Michel Simon, Les ouvriers et la politique: permanence, ruptures, réalignements, 1962-2002 (Paris: Presses de Sciences Po, 2004); Henri Rey, La gauche et les classes populaires (Paris: La Découverte, 2004).
  • [53]
    Cf. Guy Michelat and Michel Simon, “Déterminations socio-économiques, organisations symboliques et comportement électoral”, Revue française de sociologie, 26, 1985, 50.
  • [54]
    This expression is used to mean: “once we have taken account of the variables entered into the model”.
  • [55]
    This division into sextiles creates six income categories with equal number of cases in each (roughly 16.7% each).
  • [56]
    34% of those in the sample live below the poverty line, 66% above. Logarithmic transformation is useful when differences on a linear scale do not have the same sociological significance at the top or bottom of a continuum; for example, an income that increases from 6,000 to 6,500 euros does not have the same implications as one that increases from 500 to 1,000 euros.
  • [57]
    The tables in the appendix illustrate the fact that, for all of the dependent variables, the effects of assets and income do not disappear (the odds ratios are strong).
  • [58]
    On the effect of employment status on the vote in the 2007 presidential election, see Élisabeth Dupoirier, “Le parti socialiste et la gauche: l’implacable spirale de l’échec”, in Pascal Perrineau (ed.), Le vote de rupture: les élections présidentielle et législatives d’avril-juin 2007 (Paris: Presses de Sciences Po, 2008), 143-72. Dupoirier observes that 71% of self-employed people voted for Sarkozy during the second round, as opposed to 54% of private sector employees and only 42% of public sector employees.
  • [59]
    Translator’s note: Ref. means that the row is the reference case; the numerical value is left out to avoid perfect co-linearity with the other categories.
  • [60]
    Translator’s note: BEPC = Brevet d’études du premier cycle, a qualification awarded to students at the end of troisième (normally at age 15; U.S. equivalent of ninth grade).
  • [61]
    Ralf Dahrendorf, Class and Class Conflict in Industrial Society (Stanford: Stanford UP, 1959).
  • [62]
    We would like to thank Bruno Cautrès and Flora Chanvril for their comments and plentiful advice.
  • [63]
    Translator’s note: Bac+2 means that an individual has had two years of study after passing the baccalaureate exam, equivalent to an Associate’s degree in the US.
  • [64]
    Worker attributes were determined using the following process: since we already knew the profession or former profession of the respondent and if the father and mother were working when he was 14 years old, we tallied how many of the three were called a “worker” by the respondent.
  • [65]
    Since the interaction between income and assets is never significant, the set of odds ratios presented here is based on logistic regression without controlling for the effects of interaction.


Thirty years ago, the “asset”, or “property ownership”, effect discovered by Jacques Capdevielle and Élisabeth Dupoirier aroused the interest of political scientists in France and abroad in what was viewed as a heuristic subject. Since then, the “asset” variable has been almost completely forgotten. Using 2007 French presidential election data, this article demonstrates that the presence of an asset strategy remains a key factor in determining how an individual relates to politics and voting. Owning various kinds of property tends to blur the income-based distinction between “poor and wealthy” people and to neutralize the political norms embraced by the social, cultural or professional groups to which people belong. The asset effect persists when the effects of other social variables are controlled for by logistic regressions.

Viviane Le Hay
Is a CNRS-affiliated research officer at the Centre Emile Durkheim – Sciences Po Bordeaux. Her most recent publications include: “Le panel électoral français 2007: enjeux de méthode”, in Bruno Cautrès and Anne Muxel (eds), Comment les électeurs font-ils leur choix? Le panel électoral français 2007 (Paris: Presses de Sciences Po), 259-84; (with Flora Chanvril, Anne-Sophie Cousteaux, Laurent Lesnard, Chloé Méchinaud, Nicolas Sauger) “La parentalité en Europe. Analyse séquentielle des trajectoires d’entrée dans l’âge adulte à partir de l’Enquête sociale européenne”, CNAF, dossier d’étude no. 122, November 2009; (with Jean Chiche and Brigitte Le Roux) “L ’espace politique français à la veille de l’élection présidentielle (2007): analyse géométrique des données”, in Phillippe Guilbert, David Haziza, Anne Ruiz-Gazen, Yves Tillé (eds), Méthodes de sondage: cours et cas pratiques (Paris: Dunod, 2008), 152-64; (with Flora Chanvril) “Que signifie ‘‘être adulte’ pour les Français et les Européens? Des conceptions différenciées concernant les hommes et les femmes?”, European Social Survey (ESS), Paris, December 2008 ( She works on quantitative methodology and on designing questionnaire surveys in the social sciences. She is also a member of the French ESS team and teaches statistics and model design at the Centre Emile Durkheim, Sciences Po Bordeaux, (11 allée Ausone, 33607 PESSAC,
Mariette Sineau
Is a CNRS-affiliated senior researcher at the Centre de recherches politiques de Sciences Po (Cevipof). Her most recent publications include: Femmes et pouvoir sous la Ve République. De l’exclusion à l’entrée dans la course présidentielle (Paris: Presses de Sciences Po, 2011); La force du nombre. Femmes et démocratie présidentielle (La Tour d’Aigues: L ’Aube poche, 2010); “La parité législative en France, 2002-2007. Les stratégies partisanes de contournement de la loi”, Revue suisse de science politique, 14(4), 2008, 741-66; (co-edited with Manon Tremblay, Thanh-Huyen Ballmer-Cao, Bérangère Marques-Pereira) Genre, Citoyenneté et représentation, (Sainte-Foy: Presses de l’Université Laval, 2007); “Effets de genre, effets de génération? Le vote hommes/femmes à l’élection présidentielle de 2007”, Revue française de science politique, 57(3-4), 2007, 353-69; (with Vincent Tiberj) “Candidats et députés français en 2002. Une approche sociale de la représentation”, Revue française de science politique, 57(2), 2007, 163-85; “Paradoxes of the Gender Gap in France”, French Politics, Culture and Society, 24(1), 2006, 40-58. Her main interests lie in the different relationships that women and men have with politics and power. She also works on voting behavior and political behaviors according to gender, on political elites, and parity in politics. Her most recent research focuses on the relationship between wealth and politics, as well as with political trust (Cevipof, 98 rue de l’Université, 75007 Paris
Translated from French by 
James Terry
This is the latest publication of the author on cairn.
This is the latest publication of the author on cairn.
Uploaded on on 03/03/2014
Distribution électronique pour Presses de Sciences Po © Presses de Sciences Po. Tous droits réservés pour tous pays. Il est interdit, sauf accord préalable et écrit de l’éditeur, de reproduire (notamment par photocopie) partiellement ou totalement le présent article, de le stocker dans une banque de données ou de le communiquer au public sous quelque forme et de quelque manière que ce soit.
Loading... Please wait