Financial, Job and Health Satisfaction: A Comparative Approach on Working People

Abstract

The determinants of domain satisfactions could be differently evaluated depending on the aspect of life considered, which would lead to different implications for public policies. To test this hypothesis, using the German Socio−Economic Panel (GSOEP), we analyse the effect of different economic and non−economic factors on satisfaction with financial situation, job and health status. The main results confirm that several determinants exert different effects depending on the aspect of life that people are evaluating. For instance, household income only improves satisfaction with financial situation but it does not explain job or health satisfaction. However, those people with an active social life, who are less worried and distrustful, are more satisfied regardless of the aspect of life considered. These findings reflect the importance of studying the main determinants of the domain satisfactions using a comparative perspective to design and evaluate specific public policies, since some measures could be effective for improving satisfaction in one area of an individual’s life but not for others.

1. Introduction

The literature related with subjective well−being has demonstrated that it is a multidimensional concept encompassing different areas of life called domain satisfactions, that is, subjective well−being can be seen as an aggregate of different domains (for more detail, see [1,2,3,4]). Moreover, people are able to differentiate the domains and to evaluate them separately. Previous studies have considered domain satisfactions as different areas of individual life, such as the financial situation, job, health status, housing, leisure, environment, marriage, friendships, safety and social relationships [3,5,6,7,8,9,10,11,12].

Given the relationship between subjective well−being and the different domain satisfactions, the study of domain satisfactions and their determinants is also useful for policy making. Knowing what produces satisfaction in different areas of individual life could be fundamental for measuring consumer preferences and social welfare, as well as for the design and assessment of public policies [12,13,14,15,16,17,18].

Especially, the knowledge obtained in this research can be used to complement traditional measures of welfare, since the subjective vision provides information of non−material aspects of people’s satisfaction [6,19,20]. For instance, the analysis of job satisfaction and health satisfaction should be relevant for public policies related to the labour market, health care and medical expenditure.

Several studies have analysed the main determinants of different domains of life but the comparisons between the common determinants and the possible implications for public policies using this comparison are scarcer. Thus, as something new and given the relevance for the public policies, our main goal is to do a comparative analysis about the three most relevant domains, namely financial, job and health satisfaction, as well as, we focus on the common and specific implications for public policies. Specifically, using the German Socio−Economic Panel (GSOEP) over the period 1998–2014, we compare the effect of common factors, that is, those which are used to analyse the three different domains, to determine whether these affect the domains in a similar way or, in the opposite case, whether it depends on the aspect of life that people are evaluating. Moreover, we analyse the effect of several specific factors, which are included to analyse each domain. Therefore, our paper contributes to the literature about satisfaction and its implications on public policies.

Our main results show that the effect of several common factors depends on the aspect of life that people consider. For instance, absolute income (own income at the moment of the interview) only exerts effects on financial satisfaction but a higher income does not affect job, neither health satisfaction. Thus, the policy marker should consider what they want to get to design specific public policies based on the results of these studies, because, for instance, a high economic growth is better for financial satisfaction but it is not relevant to improve the satisfaction of workers with their job or their health. Nonetheless, other factors such as the social contacts explain the different domains analysed in this paper in the same direction. For instance, having more social relationships, being less worried or distrustful improves the satisfaction regardless the aspect of life that they are evaluating. Hence, our evidence confirms the relevance of social contacts to improve the satisfaction with different aspects of the individual life.

The remainder of this paper is structured as follows. The literature on domain satisfactions and determinants is reviewed in Section 2. The empirical strategy is presented in Section 3. The data and variables used to analyse the domain satisfactions are explained in Section 4. The main results of our analysis are shown and discussed in Section 5. Finally, Section 6 concludes.

2. Literature Review

2.1. Domain Satisfactions

As described in previous literature, domain satisfactions refer to the individual satisfaction in different aspects of life, such as financial, job, health, housing, leisure, environment, marriage, friendships, safety, standard of living or social relationships [1,3,5,6,7,8,9,12].1

Although the domains of life considered in the literature differ and there is not agreement on which domains are conceptually preferable, it has been demonstrated that the most standard and relevant as determinants of subjective well−being are financial situation, work and health [1,8,11,20,21,22,23,24,25]. Thus, we consider financial situation, job and health status as domains of life to analyse in this study.

Financial satisfaction is related to the current level of individual satisfaction with several aspects of their financial situation. It is well known that individual financial satisfaction could have an impact on different factors such as the choice of the consumers, productivity of the job and social contacts [26]. On the other hand, job satisfaction concerns how satisfied people are with their main activity, taking into consideration different aspects of the workplace such as wages, working hours, or relationships with co−workers and employer, among others. It is also known to be a great predictor of labour market behaviours, such as mobility, worker performance and productivity, health, longevity and social illnesses [27,28,29,30,31]. Concerning health satisfaction, it is related to the satisfaction with the current health status. It has been studied by many health economists to evaluate possible effects from illnesses and medical treatments (see, for instance, [32]), being relevant to design and assess public policy related to health care and medical spending.

2.2. Common and Specific Determinants of Domain Satisfactions

We consider both predictive and hedonistic approaches to explain the three different domains considered here (for more detail, see, [11]). In this vein, first, considering that subjective well−being can be seen as an aggregated concept of domain satisfactions [1,4], we review the factors used in subjective well−being studies, which are considered common factors of all domains and represent the predictive approach. In line with the literature, we classify them into three groups: (1) income characteristics; (2) social, cultural and psychological capital; and (3) socio−economic characteristics. Secondly, we briefly examine specific determinants, which are used for each domain in related studies and encompassing the hedonistic approach.

2.2.1. Income Characteristics

As has been demonstrated in related studies, absolute income (own income in the current period) is not the only income measure which matters in order to be satisfied with general life. Indeed, literature related to subjective well−being has widely accepted the Easterlin Paradox, that is, increases in income are not always related to increases in satisfaction, which could be explained by the individual comparisons including both internal and external. We consider that domain satisfactions could also be affected by: firstly, absolute income and secondly, the comparisons that individual makes with oneself in the past in income terms, that is, internal comparison and with peers, that is, external or social comparison [16,33,34,35]. For that, we not only include absolute income but also the relative income in the sense of a measure of internal and external comparison.

Regarding internal comparisons, it has been demonstrated for subjective well−being that increases in past income may only have a transitory effect, since either people adapt to their past experiences or new aspirations appear [36]. As a consequence, they would return to the same level of satisfaction than in an initial moment after a period of adaptation [37,38,39,40]. This process is known as hedonic adaptation.

Concerning external comparisons, these refer to the fact that comparisons are made with peers belonging to a demographic group, for instance, co−workers, neighbours, friends, family members or people with similar socio−demographic characteristics (same gender, age, education, etc.). This is usually called as relative income hypothesis. People are affected by the comparison with the economic situation of those around them, normally, their reference group. Commonly, researches impose it exogenously groping people with observable and common characteristics [41]. Moreover, the social comparisons can be modelled using two different methods, namely symmetric and asymmetric, that is, a change in the reference income influences individuals’ well−being in a similar and different manner, respectively.

2.2.2. Social, Cultural and Psychological Capital

According to the scheme proposed by [42] based on the model of Sen’s capabilities [43], the related literature has distinguished between social, cultural and psychological capital. Although these factors have received increasing attention in the literature as determinants of satisfaction [44,45,46], they are less used as determinants of domain satisfactions but we consider that these could affect them. For instance, health satisfaction could improve when people have a good mood or an active social life.

Social capital has been a debated topic, but currently there is not a common definition or consensus about how to measure it [47]. [48] based on [49] defines it as “networks together with shared norms, values and understandings that facilitate cooperation within or among groups”. As [37] pointed out, social capital includes measures of a person or group of networks, personal relationships, general trust and civic participation, called relational goods. The literature has differentiated two types of social capital: bonding and bridging. The first concerns closed networks of people with relatives or friends and the latter is more formal and it implies cross−cutting ties such as associations, trade union or the attending different social and cultural events. Previous evidence has shown that people with more social relationships experiment higher levels of satisfaction [33,35,37,45].

Regarding cultural capital, it can be defined as the values and goals in the individual’s life. The literature has shown that while the objectives social and family make to people more satisfied, the effect of economic goals is less conclusive [42]. Concerning the psychological capital, following [42], we consider the personality traits related to the so−called “Big Five Indicators” (BFI), namely neuroticism, extraversion, openness, agreeableness and conscientiousness; the LOC index as an external measure of the degree of control over an individual’s life; and a reciprocity measure (positive and negative). The existing results on subjective well−being have shown that people with more extraversion, openness, agreeableness and conscientiousness, with less neuroticism, the lower LOC (they think that external circumstances only play a small role in their life), more positive reciprocity and less negative reciprocity are more satisfied [33,50].

2.2.3. Socio−Economic Characteristics

A set of socio−economic characteristics, such as gender, place of residence, age, marital status, years of education and household characteristics, is included to analyse satisfaction. Evidence has shown that, in general, females, people who live in West Germany, with a partner or who are the owner of dwelling are more satisfied (see, for instance, [1,35,37,41]). The most extended result about age is that it has a quadratic relationship with U−shape or inverted U−shape with satisfaction [11,37]. No conclusive effects have been found for years of education. While some studies have obtained negative effects due to the fact that more educated people have more aspirations and expectations [33], others have found that more educated individuals are more satisfied [51]. The presence of children and adults in the household could have positive effects [1,33,37], negative [52] or null [34,35].

2.2.4. Specific Determinants for Each Domain

For financial satisfaction, the evidence has shown that the savings and the presence of a second earner in the household exert a positive effect [1]. Variables such as working income, working hours, extra money, extra hours or the rate between the household income and working income have been included in related papers to study job satisfaction, where a larger working income, extra money and proportion between household income and working income lead to higher job satisfaction [1]. The effect of working hours is less conclusive. While [11] found that these do not affect job satisfaction, [53] stated that a reduction of working hours could have either positive effects, since it helps to work−life balance, or negative by the association with lower working income. For health satisfaction, the factors considered have been practicing sport, where a positive effect is found since those people who do more sports, they have a better health status and the frequency of visiting to the doctor, where more visits imply less satisfaction (see, for instance, [54,55]).

3. Empirical Strategy

In line with the existing literature related to subjective well−being, the empirical model for the determination of domain satisfactions can be written as follows:

$${DS}_{it}={\alpha }_0+{\alpha }_1{y}_{it}+{\alpha }_2{y}_{i,t-k}+{\alpha }_{3}f(y_{it},y_{jt})+{\alpha }_4{SC}_{it}+{\alpha }_5{CC}_{it}+{\alpha }_6{PC}_{it}+{\rho }^{\prime }{X}_{it}+{\eta }^{\prime }{Q}_{it}+{\gamma }^{\prime }{TD}_t+{\epsilon }_{it}$$

(1)

for i = 1….N, t = 1…T, where y_it denotes the absolute income; y_i,t−k is the k−periods lagged income, that is, hedonic adaptation; f(y_it,y_jt) represents the social comparisons between the i’s income (y_it) and individual j’s income (y_jt); SC_it, CC_it and PC_it are, respectively, social, cultural and psychological capital; X_it stands for a set of socio−economic characteristics; Q_it stands for a set of specific characteristics considered in each domain; TD_t includes time dummies which account yearly changes that are the same for all individuals; and ε_it the error term. Following [11], we cardinalize our dependent variables and, then, to make use of the panel structure of the dataset, we estimate random effects model with Mundlak’s correction to control for individual heterogeneity for each domain (see, for instance, [52]). Therefore, first, we cardinalize the reported answers about the different domain satisfactions to account for the fact that pass differences among categories of satisfaction may not have the same meaning [35]. And secondly, the error term is assumed to be ${\epsilon }_{it}={\lambda }_i{\overline{z}}_i+{\omega }_i+{\pi }_{it}$, where ${\lambda }_i{\overline{z}}_i+{\omega }_i$ is Mundlak’s correction and ${{\displaystyle \pi }}_{it}$ the error term, with, ${\omega }_i\sim N(0,{\sigma }_{\omega }^2)$, ${{\displaystyle \pi }}_{it}$~N(0,1), and Cov$\left({{\displaystyle \omega }}_i,{{\displaystyle \pi }}_{it}\right)$ = 0. The Mundlak variables (${\overline{z}}_i$) used in this work are time−average values of years of education and number of adults and children in household.

4. Data and Variables

4.1. Data

The empirical analysis of this study is based on the data from the German Socio−Economic Panel (GSOEP) over the period 1998−2014. The main reasons for choosing GSOEP are its longitudinal structure and the inclusion of private households’ data to study the different domain satisfactions, such as hedonic adaptation, social, cultural and psychological capital and different socio−economic characteristics and specific aspects. To avoid the duplication of observations, we consider the responses of the household head, that is, the responses of the household member with better knowledge of the conditions in the household. Also, as [34], to control for potential panel, we consider people with three or more interviews as a proxy for the interviewing experience in the panel. Moreover, we only consider people with consecutive observations. Note that for people who are not working, there is not information on job satisfaction. Hence, to compare the different domains, we only take the specific subsample of employed people. The final number of observations is 29,430. Specifically, there are 5063 individuals, of which 32% are women and 22% of them are living in the East of Germany.

4.2. Variables

4.2.1. Dependent Variables

In the GSOEP the respondents can distinguish several aspects of life, which can be evaluated separately in terms of how satisfied people are with respect to each domain. In particular, we study financial, job and health satisfaction.

Different questions in the GSOEP about the degree of satisfaction with each domain are approximately the same “How satisfied are you with your (financial, job, health,) situation?” measured on an 11−point scale ranging from 0 (completely dissatisfied) to 10 (completely satisfied). Domains are denoted by Financial Satisfaction (FS), Job Satisfaction (JS) and Health Satisfaction (HS).

Table 1. Pearson’s correlation across domain satisfactions and general satisfaction.

	General	Financial	Job	Health
General	1.000
Financial	0.496	1.000
	(0.000)
Job	0.473	0.429	1.000
	(0.000)	(0.000)
Health	0.508	0.344	0.443	1.000
	(0.000)	(0.000)	(0.000)

Note: These are the pairwise correlation coefficients between the domain satisfactions used in this study for the whole period and general satisfaction, with p−or the whole period and gen.

reports Pearson’s correlation across the three domain satisfactions considered in this study plus general satisfaction. As in [3,8,25], all correlations are positive and statistically significant but they are not relatively high. Job and health satisfaction report a 0.443 coefficient (the highest), while health and financial show a 0.344 coefficient (the smallest). In line with previous studies, the correlation between general satisfaction and domain satisfactions is also positive, where the highest correlation is found for health satisfaction (0.508) and the smallest for job satisfaction (0.473).

Moreover,

Table 2. Descriptive statistics of domain satisfactions and explanatory variables.

Variables	Mean	SD	Min	Max
Financial Satisfaction	6.632	1.891	0	10
Job Satisfaction	6.938	1.841	0	10
HealthSatisfaction	6.757	1.876	0	10
Income Characteristics
Absolute income(a)	19.18	8529	1.135	130.1
Absolute income(a)(b)	23.74	15.16	0.200	51.48
Adaptation(a)	18.20	8.447	1.417	306.9
Adaptation(a)(b)	26.09	17.25	0.238	51.48
Poorer	0.156	0.219	0	2.77
Richer	0.122	0.200	0	1.81
Poorer(b)	0.271	0.399	0	5.129
Richer (b)	0.169	0.278	0	2.724
Social Capital
Bonding	0.431	0.495	0	1
Bridging	0.378	0.157	0	1
Cultural Capital
Eco_ goals	0.651	0.154	0	1
Fam_goals	0.819	0.202	0	1
Soc_goals	0.545	0.139	0	1
Worries	0.553	0.235	0	1
Mistrust	0.527	0.177	0	1
Risk	4.751	2.079	0	10
Psychological Capital
Neuroticism	3.724	1.150	1	7
Extraversion	4.766	1.111	1	7
Openness	4.469	1.116	1	7
Agreeableness	5.308	0.951	1	7
Conscientiousness	5.936	0.844	1	7
LOC	3.567	0.889	1	7
Positive_Rep	5.883	0.864	1	7
Negative_Rep	3.144	1.392	1	7
Socio−Economic Characteristics
Male	0.675	0.468	0	1
East	0.213	0.409	0	1
Age	45	9.340	21	74
Living _partner	0.658	0.474	0	1
Number_children	0.648	0.929	0	9
Number_adults	2.102	0.815	1	7
Years _education	12.82	2.771	7	18
Owner_dwelling	0.557	0.497	0	1
Specific Variables
Financial
Secondearner	0.892	0.310	0	1
Job
Unemployment experience	0.401	1.030	0	23
Working hours	41	8.772	1.5	80
Equivalent_extra_money(a)	25.34	42.07	0.455	145.06
Prop. Household_inc/working_inc	1.059	1.241	0.061	60
Health
Visits_doctor	8.300	13.30	0	396
Sport	3.242	1.345	1	5

Note:^a These variables are measured in hundreds of euros. ^b These variables are built considering working income rather than household income. Adapted from the German Socio−Economic Panel.

shows the main descriptive statistics of these dependent variables and of the explanatory variables whose definitions are presented in the following sections. We observe that working people report the highest average of satisfaction with their job situation and the lowest one with their financial situation (6.94 and 6.63, respectively).

4.2.2. Income Characteristics

Following [34], absolute income (y_it in Equation (1)) is obtained using household income, except in job satisfaction. This provides a measure of the more regular income components received by all household members. In the particular case of job satisfaction, we obtain absolute income using the working income.2 All income measures are real and converted into Euros for the year 2011 using consumer price index (CPI). Additionally, to control economics of scale, we calculate the equivalent income using the OECD−modified equivalence scale. We denote it as Absolute income.

We control the adaptation process including one’s own past income (y_i,t−k in Equation (1)). Although different periods have been considered in related studies, given that we do not have the same number of past observations for all individuals, we have decided to consider the lags three incomes in order not to lose a lot of observations.3 This variable is denoted as Adaptation.

Concerning external comparisons (f(y_it,y_jt) in Equation (1)), following Ferrer−i−Carbonell (2005), first, we built the reference group by grouping together all people with a similar education level, in the same age bracket and of the same region.4 Secondly, we distinguish between upward and down comparisons considering whether the individual’s absolute income is higher or lower than the average income of the reference group. Particularly, we define Poorer, when the individual absolute income is lower than the average reference income and Richer, when the individual absolute income is higher than the average reference income. They are specified as follows:

$$Poorer=\{\begin{array}{ll}{{\displaystyle \overline{y}}}_t-{{\displaystyle y}}_{it}& if{y}_{it}<{{\displaystyle \overline{y}}}_t\\ 0& if{{\displaystyle y}}_{it}\ge {{\displaystyle \overline{y}}}_t\end{array}\mathrm{and}Richer=\{\begin{array}{ll}{{\displaystyle y}}_{it}-{{\displaystyle \overline{y}}}_t& if{y}_{it}>{{\displaystyle \overline{y}}}_t\\ 0& if{{\displaystyle y}}_{it}\le {{\displaystyle \overline{y}}}_t\end{array}$$

(2)

where y_it is the individual absolute income and ${\overline{y}}_t$ is the average income of the reference group to which he/she belongs.

4.2.3. Social, Cultural and Psychological Capital

Concerning social capital (SC_it in Equation (1)), in line with the literature related to subjective well−being, we distinguish two different dimensions: bonding and bridging social capital. The respondents are asked about the frequency with which they meet with relatives and friends and their participation in different type of events, where the answers to all these questions take values between 1 “every day” and 5 “never”. We consider the categorical variable Bonding, which takes the value of 1 if the respondent meets with relatives and friends at least once a month. Bridging social capital is a linear index built using individual’s answer relating to the attendance to different types of events. Following [35], we recode the variables used to obtain bridging social capital and then, a principal components analysis is used to get the variable Bridging which is standardized between 0 and 1.

Regarding cultural capital (CC_it in Equation (1)), in accordance with the related literature, we consider three life goals: economic (success at work, having a home and affording things), family (importance of having a partner or children), and social (helping others, being fulfilled, having good relationships with friends, travel or political activity). In this case, every question is of the type “Importance of” and answers take values into the scale from 1 “very important” to 4 “unimportant”. Again, recoding this scale and using a principal component analysis, we get the normalized variables Eco_goals, Fam_goals and Soc_ goals.

Following [33], we also consider a group of variables which reflect whether people are concerned about different aspects, such as economic development, finances, peace and the environment. These variables take values between 1 “very concerned” and 3 “not concerned at all”. As in [35], we rearrange this scale and we use a principal component analysis to build the standardized variable Worries. We also take into account a variable about the mistrust of people, where the answers take values between 1 “totally agree” and 4 “totally disagree”. Using the same procedure than above, we obtain the variable Mistrust. Additionally, we include the variable Risk, which reflects the risk attitudes and takes values between 0 means the lowest risk willingness and 10 means the highest risk willingness. It is standardized to take mean zero and unit variance.

We include personal traits as part of psychological capital (PC_it in Equation (1)). Following [50], we include the BFI (Neuroticism, Extraversion, Openness, Agreeableness and Conscientiousness), the LOC index to capture the degree of control over their own life and the positive (Positive_Rep) and negative (Negative_Rep) reciprocity with others. The BFI have been obtained aggregating a total of 15 items included in the GSOEP. The external LOC is obtained after aggregating six items. Reciprocity measures, both negative and positive, are modelled by aggregation across three items each one. All these variables take values between 1 “does not apply”, and 7 “does apply”, that is, if people consider that they enclose that personal trait. Also, to facilitate the interpretation of the results, all these measures are standardized to take mean zero and variance 1.

The information of all these variables was not collected every year in the GSOEP but following [35], we impute the values for the missing year with the immediately preceding year with information and, when this is the first year, we replace it with the first data available.

4.2.4. Socio−Economic Characteristics

We use the socio−economic characteristics which are commonly considered in previous studies (X_it in Equation (1)). We define the dummy variable Male, which is coded with 1 if the respondent is man. The variable East takes the value of 1 when the respondent lives in East of Germany. The age of the respondent is included with the variable Age. To test the non-linearity in the relationship between age and domain satisfactions, we also include age squared, which is denoted as Age2. The dummy variable Living_partner takes the value of 1 if the respondent is currently living with his/her partner. We include information related to the number of children and adults in the household, which are denoted as Number_children and Number_adults. The variable Years_education measures the number of years of formal education. We also incorporate the dummy variable Owner_dwelling which takes the value of 1 if the respondent currently owns a dwelling.

4.2.5. Specific Variables for Each Domain

We consider the specific variables for each domain that have been previously used in related studies (Q_it in Equation (1)). For financial satisfaction, we consider the dummy variable Second earner which takes the value of 1 if there is more than one earner in the household. For job satisfaction, we include Unemployment experience which measures the number of years of unemployment in the respondent’s career up to the point of the interview. We also include Working hours measured as the average number of hours worked weekly. The variable Equivalent_extra_money is the sum of extra working income, including Christmas bonus, holiday bonus, 13th and 14th month and profit−sharing. It is real and converted in Euros for the year 2011. Moreover, it is corrected whit the OECD−modified equivalence scale to control the economies of scale and we consider it in logarithmic form. We also consider the ratio of household income over working income (Prop_Household_inc/working_inc). For the analysis of health satisfaction we incorporate the variable Visits_doctor which is referred to the number of visits to the doctor during the previous year and a variable about the frequency of participating in sports, which takes values between 1 “daily”and 5 “never”. Recoding this scale we obtain the variable Sport, which is standardized to take mean zero and variance 1.

5. Results

In

Table 3. Estimation results for domain satisfactions of German citizens, 1998–2014.

	FS	JS	HS
Common Variables
IncomeCharacteristics a
Absolute income	6.065 ***	0.491	0.526
	(0.628)	(0.634)	(0.696)
Adaptation	0.798 ***	−0.447 ***	0.424 **
	(0.155)	(0.119)	(0.172)
Poorer	−0.114 *	−0.011	−0.015
	(0.063)	(0.061)	(0.069)
Richer	0.084	0.139 **	0.022
	(0.066)	(0.062)	(0.073)
Social Capital
Bonding	−0.002	0.018 *	0.037 ***
	(0.009)	(0.010)	(0.010)
Bridging	0.159 ***	0.071 *	0.171 ***
	(0.037)	(0.040)	(0.046)
Cultural Capital
Eco_ goals	−0.018	0.340 ***	0.122 **
	(0.036)	(0.039)	(0.040)
Fam_ goals	0.025	0.075 **	0.025
	(0.029)	(0.031)	(0.032)
Soc_ goals	0.029	−0.050	0.056
	(0.041)	(0.044)	(0.045)
Worries	−0.458 ***	−0.341 ***	−0.212 ***
	(0.020)	(0.022)	(0.022)
Mistrust	−0.242 ***	−0.303 ***	−0.180 ***
	(0.033)	(0.036)	(0.037)
Risk	−0.003	0.014 **	0.017 **
	(0.005)	(0.006)	(0.006)
Psychological Capital
Neuroticism	−0.031 ***	−0.071 ***	−0.093 ***
	(0.007)	(0.007)	(0.007)
Extraversion	−0.002	0.003	−0.008
	(0.007)	(0.007)	(0.008)
Openness	0.007	0.013 *	0.019 **
	(0.007)	(0.007)	(0.008)
Agreeableness	0.020 **	0.036 ***	0.042 ***
	(0.007)	(0.008)	(0.008)
Conscientiousness	0.025 ***	0.050 ***	0.038 ***
	(0.006)	(0.007)	(0.007)
LOC	−0.061 ***	−0.051 ***	−0.039 ***
	(0.007)	(0.007)	(0.008)
Positive_Rep	0.035 ***	0.020 **	0.006
	(0.006)	(0.007)	(0.007)
Negative_Rep	−0.017 **	−0.023 **	0.001
	(0.007)	(0.007)	(0.008)
Socio−Economic Characteristics
Male	0.003	0.015	−0.016
	(0.019)	(0.022)	(0.021)
East	−0.152 ***	−0.040	−0.087 **
	(0.024)	(0.030)	(0.027)
Age	−0.268 ***	−0.296 ***	−0.290 ***
	(0.052)	(0.055)	(0.057)
Age2	0.275 ***	0.293 ***	0.155 **
	(0.055)	(0.059)	(0.061)
Living_partner	0.063 ***	−0.033 *	−0.011
	(0.016)	(0.017)	(0.017)
Number_children	0.073 ***	0.019 *	0.004
	(0.009)	(0.010)	(0.010)
Number_adults	0.092 ***	0.025 **	0.012
	(0.008)	(0.010)	(0.009)
Years_education	0.136	0.011	0.119
	(0.100)	(0.108)	(0.111)
	FS	JS	HS
Owner_dwelling	0.047 ***	0.013	0.011
	(0.013)	(0.014)	(0.015)
Specific Variables
Financial
Second earner	0.032 *
	(0.018)
Job
Unemployment experience		0.136 *
		(0.072)
Working hours		−0.022 **
		(0.007)
Equivalent_extra_money		0.194 **
		(0.063)
Prop. Household_inc/working_inc		0.042 **
		(0.019)
Health
Visits_doctor			−1.340 ***
			(0.032)
Sport			0.028 ***
			(0.007)
Constant	−4.698 ***	0.249	0.328
	(0.410)	(0.435)	(0.454)
Mundlak’s correction	Yes	Yes	Yes
Year dummies	Yes	Yes	Yes
Number of observations	29430	29430	29430
R−squared	0.288	0.117	0.204
Sigma_u	0.523	0.539	0.580
Sigma_e	0.537	0.590	0.594
Rho	0.487	0.455	0.488

Note: random effects estimation with standard errors in parenthesis. Columns 2−4 show the estimation for each domain satisfaction. ^a The income characteristics variables are built using working income rather than household income in job satisfaction analysis. * p < 0.1, ** p < 0.05, *** p < 0.001.

, we present the estimated results for domain satisfactions. For the sake of simplicity, we omit the estimated coefficients of time dummies and Mundlak’s correction from the table.5

Concerning common determinants, we find that the majority of estimated results of each domain satisfaction are as expected. Additionally, we observe that they exert a differential effect between domains, except social capital, being worried, distrustful, neurotic, complacent, conscientious, the Loc index related to the degree of control over the life, the gender, age and years of education. Particularly, concerning income characteristics, Absolute income only affects and positively, financial satisfaction (6.065 ***). A higher past income has positive effects on the current satisfaction with financial situation (0.798 ***) and health (0.424 ***) but this negatively affects the current job satisfaction (−0.447 ***). Therefore, the adaptation is not complete. Likewise, the asymmetric comparisons effects are confirmed, except in health satisfaction, where the social comparisons in income terms are not relevant. Our evidence shows that while having a lower income than the reference income is negative for financial satisfaction (−0.114 *), having a higher income than the average−income of the reference group is not relevant for financial satisfaction (0.084), but when working people have a working income higher than the reference income, they are more satisfied with their labour situation (0.139 **).

In terms of the influence of social capital, we find that working people with more social contacts who are less worried or distrustful are more satisfied regardless the area of life analysed. Thus, we can also confirm the relevance of the social contacts to be more satisfied, not only with the general life, as it has been demonstrated in previous studies (see, for instance, [33,35,37]) but also with these different aspects of the individual life. In terms of cultural capital, we would like to stress that the three life goals together do not affect these domains. Specifically, only the economic and/or family goals are relevant, where those people who attach more importance to economic goals are more satisfied with their job (0.340 ***) and health (0.122 **), being not significant for financial satisfaction (−0.018). People who attach more importance to family goals are more satisfied with their job situation (0.075 **). Likewise, more risky people are more satisfied with their job (0.014 **) and health (0.017 **). Relating psychological capital, we observe that, while agreeable and conscientious people are more satisfied, being neurotic exerts a negative effect on satisfaction. Moreover, when people think that external circumstances only play a small role in their life (lower LOC) they are also more satisfied.

Regarding the influence of socio−economic characteristics, we observe that there are not differences by gender on the satisfaction with these domains. The area where people live is relevant for financial (−0.152 ***) and health (−0.087 **) satisfaction, where people who are living in the East of Germany report lower levels of satisfaction, which confirm that there are differences between East and West of Germany. A U−shape relationship between age and domain satisfactions is found. Living with a partner enhances financial satisfaction (0.063 ***) but worsen job satisfaction (−0.033 *). The presence of children and adults in the households makes people more satisfied with their financial (0.073 *** and 0.092 ***, respectively) and job situation (0.019 * and 0.025 **, respectively), being no−relevant on health satisfaction (0.004 and 0.012, respectively). We find that years of education are not relevant for these domain satisfactions. Additionally, being owner dwelling only leads to more satisfaction with their financial situation (0.047 ***).

Concerning the influence of the specific variables, we find similar results to previous studies (see, for instance, [1,11]).6 Particularly, we observe that the presence of another earner in the household increases financial satisfaction (0.032 *). As [56] stated, the financial resources of two single individuals move into cohabitation change and they enjoy of a higher financial situation. For job satisfaction, the unemployment experience (0.136 *), the extra money (0.194 **) and the rate of household income over working income (0.042 **) have positive effects on job satisfaction. However, working hours are negatively associated with job satisfaction (−0.022 **). For health, higher number of visits to the doctor, which would imply that people have some health problems and a lower participation in sports decreases health satisfaction (−1.340 *** and 0.028 ***, respectively).

To conclude this section, it is worth highlighting the explanatory power of these models considering R−squared. Although we observe that it is no high, especially for job satisfaction (0.117), it is in line with the explanatory power of different models presented in the previous literature to analyse the satisfaction (see, for instance, [1,2,3]). This low R−squared reflects the complexity to construct almost a perfect model of satisfaction, in which a lot of different aspects and feelings influence people when evaluating their satisfaction with different aspects of their life, some of them inherent to their personality, or individual circumstances which are unobserved or not included in surveys to measure the satisfaction.

6. Conclusions

People are able to assess their satisfaction with different aspects of their life, known as domain satisfactions. Nonetheless, they could evaluate differently some characteristics depending on the considered aspect of life. Thus, the same factor can influence in the determination of several domains but in a different way. In this regard, the main goal of this study is to compare the determinants of satisfaction of the three main domains, namely financial, job and health situation, which give information to the governments for different decisions related to public policies.

Although, as shown in previous studies [1,3,11] domain satisfactions are interrelated due to common explanatory variables, we can observe that the same factor is differently evaluated depending on the aspect of life analysed. For instance, a higher Absolute income only affects financial satisfaction. Thus, Easterlin Paradox is only confirmed for job and health satisfaction.

Remaining an issue in the debate about whether additional indicators should complement the use of the GDP or totally replace it. As with previous studies [6,12,16,19,35,57] our evidence supports that measures of satisfaction are likely to provide additional useful information and these subjective indicators should not replace other traditional indicators but they should supplement them. It is well known that people focus on increasing their income but it does not contribute to improving neither general satisfaction nor satisfaction with other aspects of individual’s life. In this context, we conclude that this attempt to increase the income could lead to work more hours, decreasing, first, their job satisfaction and secondly, they would have less time to practice sport, increasing their stress, and, therefore, their health satisfaction would also be lower. Considering that, although economic growth is a relevant objective for governments, it does not always make people happier. However, those people who have more social contacts are more satisfied with all aspects of their life. Hence, the governments should try to create more social capital to get happier citizens and societies. For instance, they could build more meeting places as new parks. Thus, given its relevance for the general life and for different aspects of life, it can become a relevant feature of future development policies.

In addition, the evaluation of satisfaction with these areas of life is also relevant for different public policies. For instance, the assessment of the job and health satisfaction of citizens is essential to design public labour and health policies. The information on the opinion of citizens about their subjective health could be useful for the limitation of overall health and medical spending, which would improve welfare [32]. As our evidence shows, people who practice sport report higher levels of health satisfaction. Therefore, the governments could construct more places to practice sport, which would also increase the social contacts and, thus, it not only improves the health but also the satisfaction with the financial and job situation. Indeed, those more satisfied workers are more productive and healthier, which improves their performance, doing more productive the country, and, as a consequence, the health spending would be also lower.

Keeping in mind all of this, knowing which determinants influence in the domain satisfactions provides useful information to the governments to improve the satisfaction of the citizens with different aspects of life and welfare, in general. Moreover, the comparison of the effect of common variables between different domains can also help them to design more specific public policies. For instance, if they were interested in improving the job satisfaction of citizens, they could not focus on macroeconomic indicators such as economic growth, since increases in income do not enhance job satisfaction. As pointed out by [35], it is necessary to understand what really improves the satisfaction to ensure more satisfied citizens and, as pointed out by [58], happier people can live longer. Thus, analysing what contributes more happiness will facilitate happier societies and, consequently the life expectancy would increase.