Who (Really) Wins with Basic Income: Personality and Values as Predictors of Happiness Trajectories

There are many different definitions of UBI. We follow the understanding of Hoynes and Rothstein (2019), who determined three key features of UBI: a sufficient and generous cash benefit that covers a living without other earnings, independence of amount and duration of other incomes, and the availability for a wide population rather than the concentration on a particular group that tends to neediness. It is called “universal” because it is paid to all usual residents of a community, region or country. In fact, it must be acknowledged at this point that the selective and temporary design of our study lack this universal character. We decided to use the term basic income when referring to our study.

Only a few studies have investigated the influence of basic income on happiness so far. For example, one Finnish experiment on a randomized controlled trial of long-term unemployed has shown significantly higher levels of general life satisfaction as well as less distress and depression among people receiving basic income (Kangas et al., 2021). In a study conducted in Kenya, unconditional cash transfers given to the poorest of the population, increased subjective well-being, besides other outcomes (Haushofer & Shapiro, 2016). In western countries, for example in the Netherlands (Roosma, 2022), Spain (García, 2022; Sekulova et al., 2023), or Canada (Ferdosi & McDowell, 2020), projects have been restricted to samples from socially disadvantaged environments and concentrate on participants of low socioeconomic status. Therefore, their results have been shown to apply to issues of welfare assistance, but not to wider populations more generally.

Overall, income is associated with happiness (Frijters et al., 2004; Schyns, 2001; Sirgy, 2021). Utilizing longitudinal cross-lagged models, Mund et al. (2021) found strong evidence supporting a positive directional effect of income on life satisfaction, but not the other way around. Clark et al. (2008) in an economic literature review focusing on subjective well-being data from various countries, confirmed that an increase in individual income leads to an increase in well-being, up to a certain income level where subjective well-being levels off (Diener & Seligman, 2009; Helliwell, 2003; Pinquart & Sörensen, 2000). Similarly, research on well-being and wealth has suggested that a sufficient amount of money is a necessary but insufficient condition for well-being (Diener & Biswas-Diener, 2011; Kesebir & Diener, 2008). This indicates that once a moderate level of affluence is attained, income increases no longer contribute significantly to addressing personal needs. This point of satiation seems to vary across nations (Jebb et al., 2018). Conversely, inadequate financial security and resulting economic stress has been shown to pose a significant threat to happiness (Pinquart & Sörensen, 2000). For individuals experiencing this, acquiring additional income ensures access to commercial goods, social status, diverse leisure options, a sense of personal control, and unique opportunities to make meaningful contributions to society, all of which have been found to contribute to happiness (Clark et al., 2008; Diener & Biswas-Diener, 2011; Pinquart & Sörensen, 2000).

Many attempts have been made to understand what influences happiness and if it might be accomplished volitionally. The term happiness is commonly used to describe a specific pleasant state, mostly synonymous with subjective well-being (Kesebir & Diener, 2008). Subjective well-being has been found empirically to consist of two distinct components, usually referred to as cognitive and emotional well-being (Diener, 1984; Eid & Randy, 2008). The cognitive component covers the evaluation of life satisfaction, and the emotional component refers to positive and negative affect. These components have been described as related yet distinct factors for measures of subjective well-being (Arthaud-day et al., 2005; Lucas et al., 1996). Another distinction was made with the tripartite model of well-being differing between emotional (or subjective) well-being, psychological and social well-being. This has been empirically corroborated in several studies (Gallagher et al., 2009; Keyes, 2005; Kokko et al., 2013). In this understanding, emotional well-being captures happiness and life satisfaction, psychological well-being manifested as one’s attempt at self-actualization and personal growth (Ryff, 1989), and social well-being (Keyes, 1998) as one’s determination in social tasks and encounters. We refer to well-being as the emotional component and to life satisfaction as the cognitive component.

Hence, longitudinal studies typically report moderate general stability in happiness on the whole (Eid & Diener, 2004; Lucas & Donnellan, 2007). Studies that take the components separately, however, have revealed inconsistencies for happiness across the lifespan. Whereas cognitive judgements of life satisfaction have tended toward stability in adulthood and resistance to situational influence (Diener & Larsen, 1984; Diener et al., 2006; Fujita & Diener, 2005; Lucas et al., 2004), the emotional components of well-being have been revealed to be much more vulnerable to change and fluctuation (Chow et al., 2005; Diener & Larsen, 1984). A meta-analysis by Anusic and Schimmack (2016) found that, for both components, the rank-order stability with increasing time intervals shows an asymptotic course that levels off at around r = 0.30 in retest correlation.

Studies addressing the influence of major life events on happiness have shifted in their approaches over the decades. For a long time, the prevailing idea was that life events influence happiness to a certain degree, unless individuals adapt or habituate quite rapidly and return to a similar level regardless of the new circumstances (Brickman et al., 1978; Diener et al., 2006; Headey & Wearing, 1989). For example, set point theory was determining in this research context (Lykken & Tellegen, 1996). Based on the observation that people tend to revert to a baseline of happiness within a relatively short time after experiencing a positive or negative life event, a set point in life satisfaction and well-being was assumed for each individual. This was strongly supported by the Easterlin Paradox in economic life domains, i.e., the observation that people who become financially better are temporarily happier but soon habituate (Easterlin, 1974). Approaches that led in this direction were revised in their general validity by a number of studies (Diener et al., 2006; Headey et al., 2010; Luhmann et al., 2012; Yap et al., 2012). These more recent studies have shown rather that adaptation to life events might indeed proceed slowly over years, and in some cases, the process of adaptation is never completed. In addition to the research mentioned so far, studies that evaluated a variety of interventions that deliberately tried to increase well-being and mental health showed significant improvements for some intervention types (van Agteren et al., 2021).

In summary, research results have so far delivered evidence that happiness is indeed changeable. Additionally, however, evidence suggests that people adapt differently to life events and therefore vary in stability and change of their happiness, according to interindividual differences (Diener et al., 2006; Lucas, 2008; Luhmann et al., 2021).

For several decades, researchers have investigated whether there exists a happy personality. A large body of research has documented that the Big Five traits are associated with happiness. Most studies have found neuroticism and extraversion to be the key predictors for happiness (for an overview, see Larsen & Prizmic, 2008). In general, higher neuroticism is associated with lower life satisfaction, less well-being, and negative affect (Costa & McCrae, 1980; DeNeve & Cooper, 1998; Diener et al., 2003; Larsen & Ketelaar, 1991; Luhmann & Eid, 2009; Steel et al., 2008). Higher extraversion, on the other hand, relates to higher life satisfaction, more well-being, and positive affect (Costa & McCrae, 1980; DeNeve & Cooper, 1998; Lucas, 2008; Lucas & Fujita, 2000; Steel et al., 2008; Vaidya et al., 2002). In addition to the common results shown for neuroticism and extraversion, DeNeve and Cooper (1998) reported in their meta-analysis positive associations between openness to new experiences, agreeableness, and conscientiousness with various components of happiness. Conscientiousness seems to play a very important role as empirical results indicate that conscientious people generally achieve higher levels of education (Poropat, 2009), professional success (Barrick & Mount, 1991; Judge et al., 1999), higher incomes (Sutin et al., 2009), and are less likely to experience divorce, unemployment, or poor health (Roberts et al., 2007). Therefore, it is more likely for conscientious people to be satisfied with their lives (DeNeve & Cooper, 1998). Meanwhile, agreeableness and openness to new experience have proven much more ambivalent with regard to changes in well-being (Hill et al., 2014).

Less consistent research results can be derived from the relation between personality traits and happiness under changing life conditions so far (Luhmann & Eid, 2009; Yap et al., 2012). Particularly for changes in well-being accompanying shifts in income, Soto and Luhmann (2013) reported a moderating role of neuroticism on life satisfaction due to income increases. People with higher expressions in neuroticism reported even greater life satisfaction gains upon receiving an income increase than their emotional stable peers; meanwhile, those neurotic individuals also suffered more in terms of life satisfaction correspondingly due to income decreases. In contrast, Syrén et al. (2020) found no evidence for the moderating role of neuroticism in the relation between monthly income and different facets of well-being (mental, emotional and psychological), but reported mediation effects of neuroticism in the relation of income and social well-being, that is one’s determination in social tasks and encounters.

In another study by Boyce et al. (2010), conscientious people obtained more well-being in case of income increases, and on the contrary, they were suffering from unemployment in terms of life satisfaction. Therefore, we propose taking into account changes in happiness in view of the Big Five personality traits during basic income payment. Another component we consider are personal values.

Values have been defined as cognitive representations of desirable, abstract, and enduring goals which motivate actions (Roccas et al., 2002; Schwartz, 1992, 2012). The most striking distinction between traits and values concerning their content comes down to “what people are like” versus “what people consider important” (Roccas et al., 2002). While some traits and values are related to each other, they are distinct constructs (Parks-Leduc et al., 2015). Schwartz’ Theory of Basic Human Values (Schwartz, 1992) is the most influential to date and has widely confirmed ten values that can be recapped into four higher-order values (Sagiv & Schwartz, 1995; Schwartz, 2012). These are self-transcendence (universalism and benevolence), self-enhancement (achievement and power), openness to change (stimulation, self-direction, hedonism), and conservation (tradition, conformity, and security). The general idea is that people differ in activities they find satisfying or not depending on their expressions in the ten values (Oishi et al., 1999a, b). Various perspectives have emerged in accordance with the connections between these separate values and happiness (Sagiv et al., 2004, 2015), namely the healthy values point of view, value-oriented living or goal attainment, and the matching perspective.

The healthy values perspective assumes values to exist that directly enhance happiness, whereas others inhibit happiness. This view is strongly related to Self-Determination Theory. Therein, intrinsic values (e.g., finding skill development important, building strong bonds, taking on social responsibility) are oriented toward personal development and growth (Kasser, 2016; Kasser & Ryan, 1996), while external value alignment (e.g., striving for power, financial success, and status) emerges in response to negative events or when personal growth gets actively blocked (Deci & Ryan, 2000). Studies have found positive associations between components of happiness and the healthy growth-oriented values stimulation, self-direction, hedonism, and benevolence (Bobowik et al., 2011; Fetvadjiev & He, 2019; Grosz et al., 2021; Sortheix & Lönnqvist, 2014). Although the association between healthy values and well-being is mostly consistent, some studies also reported contrary results, for example, stronger reverse effects from well-being to values (Fetvadjiev & He, 2019) or no longitudinal linkages at all (Joshanloo, 2021).

A second perspective is value-oriented living and its promotion of mental health (Ostermann et al., 2017), that is, living according to one’s personal values and realizing one’s own goals. This perspective focuses on the degree to which important individual values are attained. In a German sample, Ostermann et al. (2017) distinguished between the importance and realization of personal values and implemented a measurement of realization in their study. They identified the realization of the openness-to-change values stimulation and hedonism as significant predictors for enhanced mental health and higher levels of life satisfaction. In another study, Brunstein et al. (1998) found a positive moderating effect of motive-congruent goals on well-being by differing between goal commitment and goal attainability, whereas the effect of motive incongruent goals on well-being was negative.

Besides this, the matching perspective as a third approach considers people’s personal values to fit values in the external environment and therefore suggests that congruence of both internal and external factors are crucial for well-being. On the contrary, dissonance between personal values and prevailing context causes dissatisfaction. The idea is based on the person-environment fit perspective (e.g., Kaplan, 1983; van Vianen, 2018). The evidence of matching perspective derives from different studies (Oishi et al., 1999a, b; Sagiv & Schwartz, 2000).

Hence, past research found evidence for all these perspectives, so each could be interesting in basic income context in particular. According to healthy values, there might exist some values that show associations with growth in happiness during basic income payment because they are important to people and basic income might help to easier attain (e.g., hedonism or stimulation due to greater freedom in spending behavior). There might also be values that fit the basic income context better, for example, in line with a recipient’s belief about inequalities, lifestyle, or self-determination, and therefore trajectories in happiness develop in a positive way. On the other side, some values might inhibit trajectories in happiness due to people’s desirable goals standing in contrast to what basic income implies. Although different perspectives might be useful in classifying and interpreting our findings, it is not our scope and lies beyond our possibilities to conclusively establish their plausibility, or vice versa.

To sum up, empirical studies have addressed the debate regarding associations between personality traits respectively personal values and happiness. Overall, they have led to inconsistent findings for both of our predictors. Therefore, we intend to examine personality traits and personal values and their influences on trajectories in life satisfaction and well-being in our study of basic income recipients, and we do this in an exploratory way.

The question of whether trajectories in life satisfaction and well-being relate to individuals’ personality traits or personal values deserves further investigation given the paucity of previous studies on this topic. This study aims to provide first insights into empirical data of basic income recipients and asks whether there are people who benefit (or suffer) in a special way in their trajectories of life satisfaction and well-being while receiving 1,000 Euros per month for the duration of one year as part of a basic income lottery. Based on their interindividual differences, namely on personality traits and personal values, we examine the role of these typical happiness predictors vis-à-vis basic income. We want to investigate in whether there is a relation between personality traits or personal values and happiness. More specifically, we posed two general research questions:

Do personality traits predict changes in happiness?

Do personal values predict changes in happiness?

We will control for the effects of demographics age, gender, and education. Age in general is associated with well-being (Argyle, 2001; Becker & Trautmann, 2022; Blanchflower & Oswald, 2008). Whereas some research indicates a slightly to no increase throughout lifespan (Argyle, 2001; Diener et al., 1999), new work reports a U-shape with a minimum of happiness levels in middle age (Blanchflower & Oswald, 2008). Therefore, we consider age as a covariate as it might be confounded with the dependent outcomes.

UBI is considered to favor women as it might fill the income gap and protect against poverty, especially in easing the situation of women due to the expanding opportunities to balance childcare duties and work (McKay et al., 2023). In contrast, some see women incentivized for traditional caring tasks (Robeyns, 2008). Some empirical evidence leads in the direction that an universal empowering effect on individual capacities and confidence in different areas of life can be observed (Kangas & Ylikännö, 2023). To control for this, we include gender as a covariate.

Education is considered to be linked with well-being (Diener et al., 1999; Witter et al., 1984; Yakovlev, 2012). Desjardins (2008) describes mechanisms for this. These are, at first, the direct link between education and well-being due to capabilities and resources, second the relative position of individuals in society and third that certain effects of education only materialise amongst peers with similar levels of education.

The sample consisted of 573 participants. All of them were German residents and recipients of basic income given by the crowdfunding association “Mein Grundeinkommen e. V.” (English: My Basic Income Assoc.).¹ The association ran a lottery every month. Therein, approximately 20 people were randomly selected out of a pool of voluntarily registered persons. The number of people participating in the first survey varied between nine and 46 adult recipients per cohort and was dependent on two uncontrollable events during data collection. First, the number of people who were selected relied on the previous funding of the association’s community. Second, children who won – while allowed to register online at their parents’ will – were excluded from the survey.

Each participant received 1,000 Euros per month for the duration of one year via cash transfer to their bank account. There were no exclusion criteria for the registration, although people receiving social security (named “Hartz IV” in Germany) were advised to register with a cooperating administrative body, because receiving any additional income would lower their already meager social welfare payments. Data collection was operated between October 2019 and October 2022. The sample size was only determined by the financial means that were available for the lottery, but we did not limit or enhance sample size during this period. We reported all data exclusions (not applicable), all manipulations (not applicable), and all measures in the study.

The study was designed as web-based panel study. Participants were invited from the association via email on four different occasions, of which we analyzed three waves according to the research question. These were before the payment after people were informed about the win (wave 1), in month 6 during payment (wave 2), and finally in month 12 at the time nearing the end of the payment (wave 3).

Due to the commitment of the participants to the research topic, the response rate was high with 74.1% in wave 1, 71.5% in wave 2, and 54.3% in wave 3, respectively. To further ensure a high response rate, we incentivized participation by offering a personality test with included feedback of a personal profile, made available after completion of the last survey.

Descriptive statistics for the researched variables and demographics are displayed in Table 1. Mean age was M = 39.11 (SD = 13.12) at wave 1; M = 39.59 (SD = 13.37) at wave 2; M = 41.01 (SD = 12.92) at wave 3. Gender status differed between the waves (56.4% female, 43.6% male, no nonbinary gender at wave 1; 57.6% female, 42.0% male, 0.5% nonbinary gender at wave 2; 58.0% female, 42.0% male, no nonbinary gender at wave 3). Among the participants of our study, only one person had no high school diploma at wave 1, few had absolved an apprenticeship (21.9%), while most held an academic degree (41.6%). For further descriptives, see the Supplementary Material (https://osf.io/5nhkx/?view_only=f4b21f16429c4bd5a0fb6d54bda5cba8). The raw data sample is published via https://doi.org/10.18452/28129.

Our sample is comparable to the German population. It consists of a slightly larger proportion of women (sample 56.5%, population 50.7%; Destatis, 2021) and it is younger than the average of the German population in 2021 (39.1 years vs. 44.7 years; Destatis, 2024). Furthermore, it is difficult to compare education level with official survey data for several reasons. First, data collection of the German Federal Statistical Office (Destatis, 2021) started at the age of 25 years, our sample captures people from the age of majority.

Second, we simply asked for the highest education with a mixture of general graduation and vocational qualification. In contrast, statistical surveys distinguish between these categories. At least, we could compare marginal categories with data provided by the German Federal Statistical Office (Destatis, 2021) and those who are over 25 years of age. Among these, less people of our sample have no graduation (sample 0.2%, population 4.1%), less absolved an apprenticeship (sample 21.7%, population 61.5%), more have a graduate degree (sample 43.8%, population 18.5%), and the quota of PhD is the same (both 1.3%). Due to the higher rate of graduated participants in our study, we seem to assess higher educated people of the German population.

Measurements in this study were administered in three waves. For an overview of the included items and metrics, see Appendix Table 6. For reasons of survey effectiveness we decided to give preference to commonly validated short scales. Due to language reasons, the research group translated the scale from English to German in collaboration with a bilingual expert if no German version was available.

Life satisfaction was assessed at every measurement occasion by a single item with L1 (Beierlein et al., 2014) with the question “All things considered, how satisfied are you with your life these days?” Respondents answered on an 11-point rating scale, ranging from 0 – not at all satisfied to 10 – absolutely satisfied. Mean scores ranged from M₁ = 6.16 (SD₁ = 1.85), M₂ = 6.83 (SD₂ = 1.86), to M₃ = 6.81 (SD₃ = 1.93). Rank order stability differed from r₁₂ = 0.42, r₂₃ = 0.46, to r₁₃ = 0.31.

Well-being was measured at every measurement occasion with the WHO-5 Well-Being Index (Brähler et al., 2007, for an English version, see Topp et al., 2015) with five items on a 6-point rating scale from 1 – not at any time to 6 – all the time. They included descriptions of one’s emotional state in the last two weeks (ω = 0.86 at wave 1, ω = 0.85 at wave 2, and ω = 0.87 at wave 3).

The Big Five domains on personality traits were measured in wave 1 and wave 3 following the personality dimensions outlined by Goldberg (1990) and using the German version of BFI-S by Schupp and Gerlitz (2014). The scale includes 15 items, three for each dimension of Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness. The items on the description of oneself are answered on a 6-point rating scale from 1 – does not apply at all to 6 – applies fully. Retest reliability (r₁₃) was 0.74 for neuroticism, 0.81 for extraversion and openness, 0.75 for agreeableness, and 0.67 for conscientiousness.

Personal values were assessed at every measurement occasion with a short version of Schwartz’s Portraits Value Questionnaire (Schmidt et al., 2007, for an English Version, see Schwartz, 2003). Each item was introduced with the wording “How much like you is this person?” followed by a person’s description. Answers ranged from 1 – not like me at all to 6 – very much like me. For reasons of survey length and priority of other items, we used 15 out of 21 items to assess seven of ten personal values (self-direction, power, universalism, achievement, stimulation, hedonism, benevolence). Therefore, we did not assess the higher order dimension “conservation”, which contains the scales conformity, tradition, and security. In contrast to the original version of the questionnaire, we used a gender-neutral version to address everyone regardless of gender. Retest reliability was calculated with Heise’s path analytic formula, while three time points were observed (Heise, 1969; McCrae et al., 2011) and indicated reliability estimates of 0.67 for self-direction, 0.62 for power, 0.87 for universalism, 0.70 for achievement, 0.71 for stimulation, 0.79 for hedonism, and 0.69 for benevolence.

All covariates were treated as time-invariant with measurements from the first wave. We controlled for age, gender, and education. The age of a person was assessed as continuous variable. Gender status was measured with the three categories “male,” “female,” or “nonbinary”, and education was assessed with seven categories of formal education, indicating higher education status according to the number of years spent for education.

All descriptive statistics, mean level differences, and rank-order stabilities of the assessed scales were estimated (see Table 2). Correlations among the observed variables are summarized in Appendix Table 7. To examine longitudinal changes in life satisfaction and well-being, we fit Latent Growth Curve Modeling (LGCM, McArdle & Grimm, 2010; Meredith & Tisak, 1990) in R with the lavaan package (Rosseel, 2012). LGCM estimate growth trajectories by modeling an increase or decrease over time. They consist of two latent factors reflecting the initial level represented by an intercept factor, and a change factor represented by the slope. The LGCM for the measures of life satisfaction and well-being were estimated separately. Script files can be found in the online Supplementary Material.

First, we specified changes in life satisfaction and well-being as latent outcomes. Therefore, we created a linear and a nonlinear latent growth models with fixed loadings for both the intercept and the slope. These models were assumed to reconstruct the shape of the empirical data. We tested these models against intercept-only models wherein modeling growth was omitted. After establishing the better fitting model between the intercept only model, linear growth, or nonlinear growth, we added the predictors and covariates. For purposes of interpretation, we included age as mean-centered variable in the LGCMs. Overall fit of the models was determined using the Comparative Fit Index (CFI), the Standardized Root Mean Square Residual (SRMR), and the Root Mean Square Error of Approximation (RMSEA) with 90% confidence interval. We considered models with CFI values > .95, SRMR < .08 in combination with a maximum upper bound of the RMSEA’s 90% confidence interval of < .08 to have good fit (Beauducel & Wittmann, 2005; Heene et al., 2011; Hu & Bentler, 1999). The likelihood ratio χ² statistic is also reported for the sake of completeness. Model comparison was furthermore based on the Bayesian Information Criterion (BIC) difference (Raftery, 1995). For interpreting the significance of the predictors, we set the alpha level at .01 to reduce type I errors. Further, we provide corrected p-values using False Discovery Rate procedure (FDR, Benjamini & Hochberg, 1995) to address the problem of multiple comparison.

We tested for factorial invariance to ensure that the multi-item constructs of well-being remained the same over time. We reached strict invariance for well-being (see the online Supplementary Material). We did not exclude people with missing longitudinal data, because excluding this group would lead to sample selectivity, distort the representativeness, and limit the generalizability of the results. Missing data were accommodated using Full Information Maximum Likelihood FIML; Enders, 2001), which takes all available information into account and is recommended for SEM models.

We investigated in the trajectories of happiness during basic income payment and conducted Latent Growth Curve Models to determine whether life satisfaction and well-being were changing over time. Loadings on the slope were fixed at zero for T1 and at one for T2 and T3 in the nonlinear growth models. Besides this, we tested linear growth models with loadings of zero at T1, one at T2, and two at T3, respectively. The linear growth models did not fit the data well. As a second, we included predictors and covariates of T1 into the models to estimate their predictive value on growth during basic income payment.

Overall, we did observe changes in life satisfaction and well-being. The nonlinear Latent Growth Curve Model of life satisfaction fits the data (SRMR = .033, RMSEA = .032, RMSEA 90% CI = [.000, .084], CFI = 0.99), significantly better than the intercept-only model according to the likelihood ratio test and BIC difference (χ²_Diff (df) = 62.329 (3), p < .000, Δ BIC = 43.30). The same occurred for the intercept-only and nonlinear growth models of well-being (SRMR = .035, RMSEA = .036, RMSEA 90% CI = [.000, .086], CFI = 0.97; χ²_Diff (df) = 71.944 (3), p < .000, Δ BIC = 52.90).

In the nonlinear growth model with fixed slope loadings of zero at T1, and one at T2 and T3, the average level of life satisfaction at the initial level was 6.147 (SE = 0.081, p < .001), and slope mean was 0.701 (SE = 0.093, p < .001). The intercept variance suggests significant individual differences in initial levels of life satisfaction (b = 1.567, SE = 0.277, p < .001), but no variance in growth was detected (b = 0.649, SE = 0.392, p = .098). For well-being, the initial average level was 3.431, with an increase of 0.382 (p < .001). As with life satisfaction, participants varied significantly in their well-being at the beginning of the study of the nonlinear growth model (b = 0.368, SE = 0.070, p < .001), but they did not show significant variability in the slope parameter (b = 0.202, SE = 0.103, p = .050). Notably, the correlation between intercept and slope was negative, while not statistically significant. Overall, the non-significant slope variances indicate that individuals do not differ in their growth and suggest that the limited inter-individual differences allow weak links to predictors.

In the next step, we included personality traits in the LGCM nonlinear models. Parameter estimates are shown in Table 4. Slope mean was not significant in the models indicating that no overall change had occurred. For life satisfaction, the personality traits neuroticism (b = -0.450, p < .001) as well as extraversion (b = 0.263, p < .001) predicted initial levels. Additionally, conscientiousness predicted growth in life satisfaction in the model without covariates (b = 0.267, p = .006). This was not significant when adjusting for multiple comparison (p_adj = .015). When entering the covariates into the model, the effect of conscientiousness completely diminished (p = .012, p_adj = .042). For well-being, high levels of neuroticism were associated with lower levels of well-being at the beginning of the study (b = -0.267, p < .001). For trajectories in well-being, neuroticism (b = 0.124, p = .002) and openness (b = 0.126, p = .002) were positive predictors of growth, even in the model with covariates. When controlling for multiple comparison, the p-value of neuroticism indicated no significant effect on growth in well-being, so this seems no stable predictor in the well-being model (p_adj = .014). The slopes of the personality models with covariates are depicted in Fig. 1 for the predicted factor scores of life satisfaction and in Fig. 2 for the predicted factor scores of well-being, with added Loess smooth lines.

Bild vergrößern

Bild vergrößern

The last step of the data analysis was conducted by regressing the personal values and the covariates age, gender and education on the outcomes life satisfaction and well-being. Estimates of coefficients, their standard errors, significance, and correlation between the intercept and slope parameter are displayed in Table 5.

Similar to the personality models, slope mean was not significant in the values models indicating that no overall change had occurred. For life satisfaction, the positive regression coefficients of the value hedonism (b = 0.357, p < .001) and of the covariate education (b = 0.218, p = .001) indicated that the initial level of life satisfaction was higher for people high in hedonism and for more highly educated people. Growth in life satisfaction was predicted by hedonism at T1 (b = -0.296, p = .003). This was not significant when adjusting for multiple comparison (p_adj = .011). For well-being, the initial level again was predicted by hedonism (b = 0.149, p < .001). Moreover, we detected a gender effect on the initial level of well-being (b = 0.322, p < .001), indicating that men rather than women reported having greater well-being when entering the study. In contrast to growth in life satisfaction, the personal value of self-direction predicted growth in well-being (b = 0.177, p = .003), but not when controlling for multiple comparison (p_adj = .011). This effect occurred in the value model of well-being with covariates and remained after p-value correction.

Figure 3 displays the associations between the slope of the outcome and the personal values. None of the covariates predicted growth in well-being. Furthermore, none of the correlations between intercept and slope factor across our models were statistically significant.

Bild vergrößern

Our first research question referred to changes in life satisfaction and well-being during basic income payment in relation to the Big Five personality traits. We found a predictive relation for neuroticism and openness, but conscientiousness, agreeableness and extraversion predicted no changes in happiness due to basic income. The positive effects of the two personality traits that we found can be interpreted in different ways.

First, the effect of neuroticism is in line with results from Soto and Luhmann (2013), who found an effect of income increases on life satisfaction, especially for people scoring higher in neuroticism. Additionally, we can confirm a negative relation of neuroticism to the initial level of well-being that also appeared in our sample and is well examined (e.g., Costa & McCrae, 1980; DeNeve & Cooper, 1998; Diener et al., 2003; Steel et al., 2008). This leaves more room for people high in neuroticism to enhance well-being in general. Both taken together, the lower initial level of well-being and the larger growth given higher neuroticism, we may deliver first empirical insights that the negative association of neuroticism and well-being might be softened to a certain extent during basic income payment. Underlying mechanisms for this finding are not quite clear. It might be explained by the behavioral inhibition system of neurotics because they are more sensitive to punishment and negative affect (Gray, 1970; Lucas, 2008). Soto and Luhmann (2013), concerning the results from income increases, suggest that individuals with higher neuroticism benefit from additional income in special ways because they react to negative consumption experiences more intensively. For example, this is the case when struggling to afford a living, or in regards to unfavorable social income experiences when comparing to friends with higher income, or on comparisons to one’s own past income situation after a pay cut. This explanation is supported by Trait Activation Theory by Tett and Burnett (2003), which implicates that special traits are activated by trait-relevant cues of one’s situation. Financial security offered by basic income, in this respect, might be helpful to enhance people’s emotional stability, and in turn enhance their emotional well-being, especially in otherwise challenging times. Nevertheless, these results and interpretation should be treated with caution as the effect of neuroticism on growth in well-being was not stable in the models with covariates.

Second, openness predicted growth in well-being, indicating that individuals who are open to new experiences can benefit more from receiving basic income. Our explanation behind this finding is that basic income context might offer more for people who are more open. Spending more money for pleasurable goods or events, having more room to concentrate on new experiences, autonomy and scope for action might potentially create positive affect and therefore might allow differences in openness to manifest in well-being. This explanation is underpinned by Trait Activation Theory (Tett & Burnett, 2003), which specifies that personality traits are expressed in case of trait-relevant cues, for example in specific situations or contextual settings.

Our second research question focused on the predictive character of personal values on changes in happiness during one year of basic income payment. Similar to personality traits, the trajectories of life satisfaction and well-being were predicted by different personal values.

First, we detected an increase of life satisfaction regarding lower hedonism. The growth in life satisfaction for lower levels of hedonism is in line with healthy value perspective (Deci & Ryan, 2000) as people would gain in happiness because they focus on pleasurable experiences. In other words, people low in hedonism at the beginning of basic income payment rediscover their “lust for life” and gain in overall life satisfaction. In fact, at this point a control group would be needed to statistically verify this effect. Another explanation of a more technical nature delivers the critical examination of value assessment. It is conceivable that we capture two different aspects of values previously reported by Ostermann et al. (2017), that is, the importance of values and their realization. We cannot completely exclude that some people score low in hedonism at the beginning of the study because they were not in the position to realize this value. In contrast, we were rather interested in the importance for a person. Therefore, it would be clarifying in basic income studies to assess the importance of values and realization of values to distinguish these aspects and to open new avenues for interpretation.

Second, well-being increased regarding highly valued self-direction. Self-direction is conceptualized as autonomy of thought and of action (Sagiv & Schwartz, 2000). Hence, realization of self-direction might be better fulfilled in a basic income context as there are several degrees of freedom due to added financial security. This might enable people to set new goals, discover new activities, feel more independence for example of employers, live out creativity, and therefore perceive more well-being. In line with value-oriented living perspective, basic income context would more likely enable people to live in accordance with the value self-direction, indicating an alignment of its importance for the person and realization due to financial security. At the same time, it is similar to the result for openness to new experiences reported above and Trait Activation Theory (Tett & Burnett, 2003) in that situational or contextual cues fit better to a specific value allowing for this value to be more strongly expressed.

Most effects remained stable when entering the covariates age, gender, and education in the models. In particular models, some of the covariates had an effect on the initial level of the outcome, but not on the slope. Taken together, we did not detect a particular pattern of links between the covariates and trajectories in happiness during basic income payment.

Some practical implications both for and beyond basic income can be derived from these results. Certain expressions in traits and values were associated to positive trajectories of life satisfaction and well-being. Thus, there are practical implications for emotionally instable persons, as high levels of neuroticism were associated with negative objective life events (Magnus et al., 1993), negative affect (Larsen & Prizmic, 2008), and are known to be a risk factor for mental disorders (Bagby et al., 2017). The stabilizing effect on one’s financial situation could therefore lead to much-needed relief in the reality of life for emotionally less stable persons in the long term, and thus provide overall (cost) mitigation for the healthcare system due to mental disorders (Widiger & Oltmanns, 2017). There has also been a tendency of increase in well-being for people high in openness to experience and evaluating self-direction as important. This may be related to the creative scope provided by a basic income, which people who are or want to be curious, creative, innovative thinkers and independent in particular can make use of. In the context of UBI, these people can act as role models and multipliers who break new ground and inspire others through their spirit and actions.

Here we will address a few strengths and limitations. The study presented offers several advantages in comparison to other studies on basic income. Our sample is the first of its kind in Europe due to the fact that it is not tied to any requirements concerning demographics, grade of charity, or employment status of participants. The longitudinal design, the separate analysis for life satisfaction and well-being, the relatively large sample size if one considers the costs of basic income payments per person, and the heterogeneity according to age demonstrate its outstanding quality. It replicates findings for the associations between the two personality traits neuroticism and extraversion and the initial level of happiness that highlight the comparability of our sample, but for the sake of novelty in basic income research it opens new insights into the field of happiness. Nevertheless, there certainly persists room for improvement and future investigations.

For receiving basic income and participating in the study, people had to register at a website. This might lead to self-selection effects. We earlier reported that about 40% of participants in the study had an academic degree, demonstrating some initiative toward basic income especially among more highly educated people. Additionally, we cannot check for the self-selective character regarding expressions of specific traits or values. For example, extraversion is associated with the experience of more positive life events (Magnus et al., 1993; Specht et al., 2011). People who are hedonists try to create more positive life events (Larsen & Prizmic, 2008) that might be affordable with basic income payments and stimulation, self-direction and hedonism are positively associated with the number of controllable events (Sortheix et al., 2013). Naturally, a fully randomized-controlled trial would be beneficial for future investigations.

Some small improvements we would propose are related to the study design. With regard to our research questions, we analyzed trajectories of happiness before, during, and at the end of basic income payment. As six months passed between T1 and T2, we were unable to pinpoint when the overall life satisfaction and well-being reached a plateau. To account for this more precisely, shorter time intervals would be necessary from the beginning of the study. Further, it would be quite interesting to see what happens in terms of sustainability of the observed growth. Therefore, future research should investigate happiness levels in follow up surveys. Additionally, we analyzed basic income payment periods of only one year. Now, if we were to look at basic income as an alternative security system that could be implemented indefinitely, future research would have to examine longer lasting effects. In turn, giving people a long-term perspective might evoke different reactions in people’s happiness, including habituation, recovery or stabilization. To address this hypothetical issue, experiments would have to last a minimum of several years with accompanying longitudinal research.

We did not investigate three of Schwartz’s values on the higher order dimension conservation due to reasons of survey length. As some of the values belonging to this dimension seem to be associated with conscientiousness and openness to experiences (Roccas et al., 2002), which proved to be predictors in our study, they could contribute to interpretation of changes in happiness. Neither did we investigate in changes of personality or personal values during basic income payment because such changes had not been addressed by our research questions. Nevertheless, life events might cause changes in personality (Bleidorn et al., 2018; Specht et al., 2011) and in values (Bardi et al., 2014; Milfont et al., 2016) to a certain extent. Therefore, it is necessary to investigate changes in these interindividual differences regarding additional income.

The underlying assumption of our study was the idea that people with an additional income might choose activities and practices that enhance their life satisfaction and well-being. Several ideas about changes in happiness are available. Lyubomirsky et al. (2005) have shown that distinct components influence happiness, including a volitional component. Surely, what changes happiness concretely might vary from person to person. For example, Oishi et al., (1999b) have reported that people differ in the type of activity they find satisfying depending upon their values. For the sake of understanding such mechanisms, future studies should investigate if there exists a pattern of concrete activities, decisions, or practices under basic income common to those persons holding the same traits or values.