Mental well-being and government support in Europe. The mediating role of trust in people and institutions

Purpose – This paper investigates the determinants of subjective well-being in Europe using the European Living, Working and COVID-19 (ELWC) Survey carried out by Eurofound (2021). Socio-demographics characteristics, employment status, measures of economic distress, inequality and work life balance are considered. Particular attention is paid to how quality of government support (QGS), that considers the dimensionsofgoodgovernancesuchasintegrity,fairness,reliability,responsivenessandinfluencessubjectivementalwell-being(WHO-5)throughthemediationoftrustinotherpeopleandininstitutions. Design/methodology/approach – To this end, the authors estimate a moderated mediation model for analysing the indirect role of QGS on WHO-5 through institutional trust and trust in people. Findings – TheresultssupportthehypothesisthatthereductioninWHO-5intheEuropeanpopulationduring coronavirus disease 2019 (COVID – 19), particularly marked in the 18 – 34 age group, is related to the perceived inadequacy of government interventions in managing economic and social uncertainty through supportive measures. This outcome is also due to reduced trust in institutions and other people, as both are significant mediators that reinforce the impact of public support on WHO-5. Practical implications – Government should pay greater attention to this relationship amongst good governance, trust and mental health of citizens because a healthy human capital is a significant factor for the long-run economic growth, in a special way when the authors refer to the young workforce with a greater life expectancy. Originality/value – In the literature, the role of trust as a mediator has been analysed in the relationship between individual economic situations and subjective well-being before and during the COVID-19 pandemic. Tothebestoftheauthors ’ knowledge


Introduction
The spread of coronavirus disease 2019  has resulted in an impact on multiple aspects of life, from employment and financial insecurity, to an increasing need for work-life balance and concerns about quality of life and people's subjective well-being.
Together with non-pharmaceutical intervention (NPI) aimed at stopping the spread of the virus, countries have implemented economic support measures able to positively counterbalance the negative impact of closure policies on economic growth (Alfano et al., 2022). Apart from the economic result of contextual implementation of containment and mitigation policies, states' actions to cope with health and socioeconomic crises act on a psychological level, providing guidance for behavioural responses to uncertainty, when financial insecurity is a major source of negative feelings and mental distress. The quality of public action can, therefore, contribute to restoring subjective well-being amongst the public, in addition to taking care of their economic interests during the COVID-19 pandemic. A rich branch of socio-economic literature sheds light on how subjective well-being is influenced by welfare and social interventions and, more generally, by the quality of government support (QGS) such as responsiveness and reliability, as well as the values of fairness and integrity (Helliwell and Huang, 2008;OECD, 2017b;Veenhoven, 2010). In order to reduce uncertainty in times of crisis, governments increase citizens' well-being when they promote accessible and efficient support that effectively addresses the needs and expectations of society as a whole. Fairness and openness in the process and outcomes of public policies also play a role in ensuring that supportive action is successful (Charron et al., 2015;Kaufmann et al., 2011;OECD, 2017a).
A further issue is the mechanism through which public action impacts on individual mental well-being. It may act directly or through other determinants of subjective well-being. The COVID-19 pandemic gives rise to a greater need for trust as the public authorities rely on individual compliance to make anti-contagion measures and behavioural guidelines effective. In turn, individuals need to feel that the authorities have the power to do something to protect them. Trust plays a role in the effectiveness of public interventions considering that it influences the relationship between the general public and government and, in turn, it may make public interventions more or less successful.
Some authors suggest that trust is also a key factor for public health in times of crisis. In the COVID-19 period, low trust induces socially uncooperative behaviour, increased mobility of individuals and potentially results in higher mortality growth rates (Bargain and Aminjonov, 2020). Similarly, Alfano (2022) found that European countries with higher social capital stocks have fewer COVID-19 cases, ceteris paribus as a result of government recommendations and restrictions.
Given the significance of this dimension, we enrich the analysis exploring the direct and indirect factors influencing the impact of government policies on individual well-being, investigating the mediating role of trust in people and institutions.
Specifically, we examine whether (1) the perceived QGS is successful in promoting mental well-being, (2) the QGS has an impact on trust in people and institutions, (3) trust in institutions and in other people influence subjective well-being, (4) institutional trust and trust in people play a mediating role in the relationship between the perceived quality of public support during the COVID-19 pandemic and subjective well-being and (5) this process of mediation was different for the various age groups of population. To this aim, we estimate a moderated mediation model using the database drawn from the European Living, Working and COVID-19 (Survey, carried out by Eurofound (2021). With regard to the assessment of subjective well-being, we use the index validated worldwide, the WHO-5 mental well-being index developed by the World Health Organisation (WHO), whilst to capture the perceived QGS we provide an original indicator identifying the dimensions of responsiveness and IJSE reliability of state capabilities, integrity and equity values, such as transparent and fair rules for access to income support measures.
To the best of our knowledge, no studies have examined the role of perceived QGS on subjective mental well-being using the mediating and backing effects of trust in people and institutions. Up to now, trust has been used as a mediator only in the relationship between individuals' economic status and subjective well-being before and during the COVID-19 pandemic (Clench-Aas and Holte, 2021;Lee, 2022). In addition, another novelty of the present work is the analysis of the moderating role of age in the relationship between government support, trust and subjective mental well-being, in line with the view that a particular moment in history may affect trust across age groups and generations to a different extent (Eichengreen et al., 2020). Specifically, we analyse the 18-34 age group based on statistical evidence from the Eurofound (2021), which shows that this group has experienced the greatest reduction in both mental well-being (e.g. risk of depression, perceived anxiety and stress) and economic and financial well-being, for example, due to higher rates of unemployment and job insecurity (Lambovska et al., 2021;Rossi et al., 2020).
We find a positive relationship between WHO-5 and perceived QGS (direct effect) and the presence of indirect positive effects produced by QGS on subjective mental well-being through trust in people and institutions. In addition, worsening subjective well-being conditions are partly due to reduced trust in institutions and other individuals, as both are significant mediators that enhance the positive impact of public support on well-being.
Finally, institutional trust has less of an indirect effect on WHO-5 in the younger age group (18-34 years old), which may suffer a major lack of influence in political decision-making, with negative consequences on their trust in institutions. Nonetheless, they attribute a higher relevance to the trust in other people, with respect to the individuals over 34.
The paper is structured as follows. In Section 1 we develop the theoretical background. Section 2 outlines data and methods used in the empirical strategy. The results are pictured in Section 3. Section 4 includes the discussions and Section 5 presents conclusions and policy implications.
2. Theoretical background 2.1 Responding to individual needs in time of crisis: public support and subjective well-being In recent years, many scholars have focussed on whether and how subjective well-being is influenced by welfare supports and public services. As highlighted by the OECD (2017a), good governance captures factors such as responsiveness, reliability, fairness, openness and integrity of public interventions.
The empirical literature has found that states that are characterised by low levels of transparency in the public sphere and low levels of perceived fairness in public policy are associated with lower levels of happiness (Veenhoven, 2010) and diminished overall subjective well-being (Helliwell and Huang, 2008). Helliwell (2003) compares different aspects of subjective well-being to a set of government measures using the World Bank Worldwide Governance Indicators (WGI) and finds that government effectiveness, a functioning regulatory framework, a guaranteed rule of law and robust action against corruption are closely associated with higher levels of subjective well-being.
The main insight obtained from these studies is that well-being depends more on the perceived quality of the services provided than on the funding allocated to their provision.
In this connection, Kaufmann et al. (2011, p. 239-240) claim that "first, perceptions matter because agents base their actions on their perceptions, impression, and views [. . .] Second, in many areas of governance, there are few alternatives to relying on perceptions data. For instance, this has been particularly the case for corruption, which almost by definition Mental well-being and government support leaves no 'paper trail' that can be captured by purely objective measures. Third, we note that even when objective or fact-based data are available, often such data may capture the de jure notion of laws 'on the books', which often differs substantially from the de facto reality that exists on the ground". Since these arguments appear to be convincing, in the empirical part of our paper we propose a measure of QGS based on perceptions of public sector interventions by which we seek to capture how individuals judge the responsiveness, reliability, fairness and integrity of public intervention.
2.2 Leveraging trust for well-being: trust as a key mediator Good governance can improve the evaluation of quality of life either directly, in the sense that individuals feel better living in a context of high quality of public support, or indirectly, in the sense that good governance enables individuals to achieve higher levels of other determinants of subjective well-being, including trust in institutions and in people.
To address this issue, we take the study one step further by integrating QGS, trust and subjective well-being for the purpose of examining the processes by which public policy can have a bearing on individual subjective well-being.
Many authors suggest that trust is a key factor for public health in times of crisis: Pagliaro et al. (2021), in assessing individual experience of trust, associated with the willingness to coordinate efforts with others and cooperate with requests from the authorities, find that psychological differences in terms of trust in government, people and science are a good predictor of individual behavioural responses and cooperation with government needs for compliance across countries. Similarly, Alfano and Ercolano (2021) highlight that both local social capital and institutional quality have affected the efficiency of lockdown measures in Italian provinces.
According to the Organisation for Economic Co-operation and Development (OECD) guidelines for measuring trust (2017b), the multidimensional concept of trust can be distinguished as follows: a general trust in other people, including individuals with whom no direct relationship exists but who belong to the same community and those personally known, and a specific trust in well-identified political and non-political institutions.
Since seminal Putnam's (1993) studies an overlap between the concept of trust and the concept of social capital has been emerging. Putnam himself argues that the concept of social capital «refers to features of social organisation, such as network, norms and trust, that facilitate coordination and cooperation for mutual benefit» and he defines trust as a «lubricant » for social life (Putnam, 1994, p. 7-9). Later, the same author argues that, generally speaking, the decline in institutional trust is due to the citizen's perception of a worsening in the performance of many Western institutions (Putnam, 2000).
The process of transmission by which public policies affect trust in the public institutions has been referred to in the literature as the micro-performance hypothesis: better quality of governance can result in people being more satisfied, which in turn can generate increased institutional trust ( Van de Walle and Bouckaert, 2003).
It is not surprising that responsive and effective public governance have an impact on trust in institutions (Van Ryzin, 2007). However, at the same time, a branch of empirical research shows that when a society is not governed in a way that fosters equity among all its members, trust in people declines. The traditional institution-centred argument (Rothstein and Stolle, 2008) claims that when institutions are universally oriented and act effectively and fairly, individuals tend to think that the state will successfully intervene to avoid opportunism and to safeguard individual rights. In turn, this will encourage cooperation and lower the risk perception in trusting others.

IJSE
Considering that good governance can have a positive impact on trust in people and in institutions, we should bear in mind that a society characterised by high levels of trust has an impact on mental well-being. Barrafrem et al. (2021) report that trust in government to deal with healthcare challenges arising from the COVID-19 pandemic had a significant direct impact on an index of individual well-being (Diener et al., 1985). These results reinforce other studies, such as Clench-Aas and Holte's (2021) research, which found that individual life satisfaction is directly associated not only with trust in other people, but also with trust in institutions.
With regard to the key role of trust, different recent studies investigate the mediating role of trust or social capital during the pandemic on the well-being and mental health of citizens on large data sets. In particular, Lee (2022) analyses the mediating role of trust in institutions with respect to its ability to reduce the negative effect of financial distress on mental well-being, finding that this dimension has a significant indirect impact. The author focusses on middle-aged and older adults. Otherwise, we decide to focus on the younger group of individuals, under 35 years old, adopting a moderate mediation that allows us to study the differences between the two age groups. Other studies on the buffer effect of trust include those of Chan et al. (2021) and that of Ciziceno and Travaglino (2019). The former adopts the more general dimension of social capital as a mediator between COVID-19 concerns and mental health problems. The latter shows that trust in institutions is a significant mediator between perceived corruption, which can be considered a proxy for the QGS and life satisfaction, which is a dimension of individual well-being.
The theoretical connection between trust, subjective well-being and the mediating role of trust in people and in institutions provides the ground for the transmission mechanism developed in the following empirical analysis. The path diagram theoretically discussed is provided by Figure S1, supplementary material.

Data and sources
To explore the hypotheses arising from the theoretical framework above, we employ a database drawn from the European Living, Working and COVID-19 Survey carried out by Eurofound (2021) [10] developed during the COVID-19 pandemic (ELWC). The survey was conducted in three different rounds: the first one was launched in April 2020, the second one in July 2020 and the third one was implemented in February and March 2021. The three rounds reached almost 190,000 European citizens. The aim of the ELWC is to examine changes taking place during the pandemic in terms of trust in the public institutions, working conditions, financial and economic instability and other dimensions which capture individual well-being, such as life satisfaction, positive and negative attitudes and other measures of mental well-being. The ELWC also measured the assessment of the survey respondents with respect to income support and social and financial assistance provided during the pandemic by national governments and the European Union. We selected the third round of the ELWC involving 46,800 respondents (Eurofound, 2021). This round comprises variables of interest for our study, not sampled in the previous rounds, for example, interpersonal trust.
Most of the variables we analysed are related to perceptions, such as the perception of how fair the measures are, how much individuals trust institutions and others and the perception of the individual financial situation. Subjective perceptions and impressions appear to influence agents' choices and evaluations of public measures even more than direct and personal experiences. In fact, individuals make their overall assessment on the basis of past experiences or evaluations shared by other individuals in their social networks, as well as on the basis of public information and media reporting. We return to the role of perceptions in the following sections.

Preliminary analysis and variable definition
Our sample consists of individuals from the European Union, [1] mainly women (63%), in the age range from 18 to 98 years. The majority of the sample is characterised by individuals over 50 years of age (61%), followed by individuals from 36 to 49 years (25%), whilst individuals from 18 to 34 represent a minority of the sample (14%). Most of the interviewees were employees (49%), a significant share consisted of retired people (24%), followed by the unemployed (9%), self-employed (8%), homemakers (4%), individuals unable to work due to disabilities or long-term illness (3%) and students (3%). Moreover, the majority of our sample had a high level of educational qualifications: 51% had a bachelor's degree, a master's or a doctorate. With regard to household composition, 61% of the sample reported having a spouse or partner and 45% had, at least, one child up to 11 years old. A total of 79% live in a household consisting of one to three individuals. In terms of work-life balance, from the first wave in April 2020 to the second wave in July 2020 there was a general worsening that was particularly significant for women working from home with children. However, statistics relating to the third wave, in March 2021, show a partial improvement in work-life balance conditions (Eurofound, 2021). In fact, in the sample as a whole, 60% of the respondents stated that they worried about work when not working at least sometimes, 73% felt too tired after work to do some of the household jobs that needed to be done, 55% reported that their job prevented them from dedicating the time needed for their family.
With respect to subjective mental well-being, the European Living, Working and Covid-19 Report provides evidence of a significant deterioration between the first and the third e-survey, in which the lowest levels of mental and psychological health were measured. This worsening of mental health is particularly evident in the 18-to-34 age group and for the unemployed (Eurofound, 2021). In our sample, our calculations show that the average levels of mental well-being are relatively low: the mean score on the mental well-being index, [2] ranging from 0 to 100, was equal to 47. Moreover, low levels of mental well-being were reported by 56% of the total sample, placed in the first and second percentile of the mental well-being index. In addition, 29% of respondents stated that they were downhearted and depressed more than half of the time, and 36% of respondents suffered from frequent feelings of tension, whilst 28% reported persistent feelings of loneliness and isolation. These average levels were lower for young people aged from 18 to 34 years, with an average mental wellbeing index of 44, whilst the share of young individuals who reported feeling frequently depressed or downhearted rose to 35% and those reporting frequent feelings of tension amounted to 46% of the sample. Similarly, the perception of loneliness and isolation was higher amongst young people, 35% of whom report habitually suffering from these emotional states. In terms of financial and economic difficulties, we find that almost onequarter of the total sample reported struggling to make ends meet and a worsening of their financial situation.
Another important dimension is institutional trust, which significantly declined between the first to the third round of the Eurofound survey. In particular, institutional trust increased between the first round in April 2020 and the second round in July 2020, whilst it significantly declined between the second and the third round in March 2021, falling to average levels below those of April 2020 (Eurofound, 2021). In our sample, from the data analysis, we found that the police and the healthcare system were the institutions that individuals trusted more: 47% of the sample reported higher levels of trust in the healthcare system and 44% in the police. High levels of trust in the European Union (34%) and in the government (26%) were reported by a minority of individuals. The higher levels of trust reported by a significant part of the sample to the police and the healthcare system may be linked to the type of public service offered by these institutions during the pandemic: the proximity of the service and direct contact with the public are elements that can play an important role in determining levels of trust. Moreover, the police and the healthcare system are institutions with a primary role in containing the IJSE pandemic during the state of emergency. We also analysed the levels of trust in other people, finding that only 34% of individuals reported high levels of trust in other individuals: Europeans seem to trust institutions, particularly the police and healthcare system, more than they trust other people. Only trust in government seems to be lower than trust in other people in general. Lower levels of interpersonal trust may reflect growing inequalities amongst the most vulnerable groups of the population and the characteristics of fairness and equity of the social support measures adopted during the emergency (Eurofound, 2021).
With regard to access to income support measures, the findings show that the percentage of individuals who effectively assessed them was low: 87% of respondents reported that they did not make use of income support measures such as state aid for businesses, unemployment benefits, wage support, paid sick or care leave and other forms of assistance. This provides evidence that the majority of the sample consists of individuals who evaluate the efficacy of support services without any direct experience of them. Thus, the opinions expressed about income support measures, which we consider in our analysis, reflect the general perception of the ability of the institutions to respond to social needs. With respect to those making use of income support measures, the findings show that the most significant category was state aid for businesses, requested by 34% of the total sample, unemployment benefit, requested by 9% of the sample, paid leave or sick leave (8%), other types of social support (8%) and wage support (7%).
For the sake of brevity, the intercorrelations between dependent and independent variables are not presented in the text, but these are available in the supplementary material to provide an initial picture of the strength of the relationships between the pairs of variables used in the moderated mediation model (Table S4).

Variables derived from non-linear CATPCA
In the moderated mediation model proposed in this paper, we included composite indices described by factors derived using Categorical Non-linear Component Analysis (CATPCA). An important advantage of CATPCA is its ability to deal with categorical variables by discovering and managing non-linear relationships among them, particularly in the presence of large datasets characterised by the presence of different types and scales of categorical variables, which are often correlated nonlinearly (Linting et al., 2007). Table S3, in the supplementary material, shows the Cronbach's alphas of each factor, which suggest high internal consistency among the items. It also displays that the variance accounted for (VAF) each factor is high and above 90%.
In summary, we used both the variables directly adopted from the dataset provided by Eurofound and the additional composite indices (factors) presented in Table 1, calculated through non-linear CATPCA (for details see Table S1, in the Supplementary File).
In the following, we present the variables used. Quality of Government Support (QGS): this factor comprises four items from the survey relating different dimensions of welfare policies adopted by the national government. As suggested by OECD (2017b), we consider state competencies in terms of responsiveness and reliability, as well as the values of fairness and equity, representing important dimensions influencing institutional and interpersonal trust. Specifically, this factor comprises the following variables: clarity and transparency of support measures (integrity), fairness, the ability to reach those most in need (reliability) and efficiency in terms of obtaining benefits (responsiveness). The Cronbach's α of this factor is equal to 0.902.
Institutional Trust: this is a factor computed by considering local, national and multinational dimensions. In particular, we include in this factor the level of trust by individuals in national governments, healthcare systems, the police and the European Union.

IJSE
Economic distress (ECO-STRESS): This factor combines economic and financial variables linked to past performance and expectations concerning the future financial condition of respondents. In addition, it comprises a variable that measures the degree of poverty individuals face in the present moment. The factor has a Cronbach's α equal to 0.737.
Work-Life balance (WLB): this factor is related to work-life balance, including five items concerning the ability to balance work and life spheres. The Cronbach's α is equal to 0.784.
We then consider additional variables already included in the original database provided by Eurofound (2021): WHO-5: this is a subjective composite well-being index developed by the WHO for screening depression and measuring mental well-being. The index ranges from 0 to 100, with higher scores denoting higher levels of well-being. It is derived from five statements indicating the frequency with which individuals felt, in the last two weeks, cheerful and in good spirits, calm and relaxed, active and vigorous, fresh and rested and interested in daily activities of their life. This index has been validated in studies on both younger and older persons, confirming that this scale has good construct validity (Topp et al., 2015;WHO, 1998).
Trust in people: this variable expresses how much individuals trust other people, in a range from 0 to 10, with higher scores indicating higher levels of trust in other people.
Young: this dichotomous variable is based on the threshold adopted by the Eurofound (2021) considering young Europeans aged 18-to-34 years.

Empirical strategy
We estimate a moderated mediation model with Hayes' PROCESS Macro (Version 4.0, released in August 2022) (Model 58) for analysing our hypotheses relating the indirect role of QGS on WHO-5 through Institutional Trust and Trust in people. The indirect paths are estimated with the moderator Young for assessing the presence of significant differences between the 18 and 34 age group and those over the age of 34. This is also supported by the t-tests displayed in Table S2 in the supplementary file. The moderated mediation and the path diagram ( Figure S1, supplementary material) are represented by the following equation models: Whereas x i is represented by the variable QgS, M 1i and M 2i indicate, respectively, the mediators of Institutional Trust and Trust in People, adopted as dependent variables in equation (1) and (2) and then as regressors in equation (3). The dependent variable Y i in equation (3) is identified by WHO-5; w is the moderator identified by the dummy Young. To estimate the moderation effect, we consider the interaction term ðx 3 wÞ i in equation (1) and (2) and the interaction terms ðM 1 3 wÞ i and ðM 2 3 wÞ i in equation (3). We also include controls represented by socio-demographic characteristics, variables concerning employment status, work-life balance conditions, economic distress and the perception of inequality.

Results
The results of our moderated mediation analysis may be outlined as follows. First, we show the impact of QGS on Trust in People and Institutional Trust (Models 1 and 2), and, second, we present the results relating to the impact of QGS, through Trust in People and Institutional Trust (mediators) on WHO-5 (Model 3). Coefficients between the mentioned variables are differentiated by the moderator Young. We also briefly comment on the control variable coefficients without carrying out an in-depth analysis with further comments.
4.1 Analysis of the quality of government support on institutional trust (Model 1), trust in people (Model 2) and age-related differences First, QGS was found to be a significant and positive determinant of both Institutional trust and Trust in other people (Models 1 and 2). The ability of the authorities in terms of responsiveness and reliability and respect for fairness and integrity values are attributes with a significant impact on both trust in people and institutions. The interaction (QGS 3 Young) on Institutional Trust is significant and negative, indicating that QGS was found to be a weaker determinant of institutional trust for individuals aged 18-to-34 years. This result suggests that the young people may be less involved in welfare policies and thus perceive a greater distance from the institutions, with important consequences in terms of trust in the institutions.
The 18-34 age group is also characterised by lower levels of trust in other people as supported by the coefficient of the dichotomous variable Young (in Model 2). However, the interaction (QGS 3 Young) shows a positive and significant impact on Trust in People (Model 2). Although young people tend to trust other people less, QGS has a higher positive impact on Trust in People in this age group than for the over 34 age group. This result may suggest that young Europeans assign more importance to issues of fairness, transparency and integrity that have a significant impact on trust in other people.
4.2 Implications of institutional trust and trust in people on subjective mental well-being, mediation effects and age-related differences (Model 3) With regard to the subjective mental well-being model, the findings show that the QGS has a positive and significant impact on WHO-5: individuals who positively evaluate the overall quality of governance show higher levels of WHO-5 (Model 3). Further, institutional trust and trust in people are also dimensions that enhance individual well-being. Institutional Trust, which has a positive and significant coefficient, helps individuals to deal with economic and financial risks, whilst increasing the individuals' perception of control over their lives (Barrafrem et al., 2021;Roccato et al., 2021). Similarly, Trust in People is another positive and significant determinant of WHO-5.
The index of Institutional Trust is equal to À0.523, CDI 5 [À0.812; À0.229], indicating that the indirect positive effect on WHO-5 produced through trust in institutions is less significant for the young population. On the other hand, the index of moderated mediation of Trust in People is equal to 0.149, CDI 5 [0.015; 0.293], suggesting that the indirect positive effect generated by trust in others is higher for young individuals. We examine these different paths in greater depth below.

The impact of control variables and socio-demographic characteristics on well-being (Model 3)
Women experience lower mental well-being than men. Conversely, having at least one minor child (<12 years old), having a partner and good levels of work-life balance are all dimensions positively related to WHO-5. Individuals who are subject to conditions of economic and IJSE financial instability present lower levels of subjective mental well-being. With respect to the occupational dummies, the retired and self-employed represent the job occupation with the highest WHO-5. The students, who represent the only work categories not considered in the model, are one of the categories with the lowest WHO-5 together with the unemployed. For reasons of brevity, we outline the control variables only in this paragraph, without carrying out an in-depth analysis with further comments.

Direct and indirect paths of trust in people and institutions
In the following, we compare the positive indirect effects of institutional trust and trust in other people with the total effect of the QGS on WHO-5. Moreover, we differentiate the indirect effects of Trust in People and Institutional Trust on the 18-34 age group and those over 34 (see Table 2).
By comparing the coefficients of the indirect effects of Institutional Trust and Trust in People with the total effect, the results show that these two indirect channels of trust account for more than 50% of the total effect produced by the QGS, in both age groups. This result shows that the positive effect of government support on well-being is in a large part determined by its effect of increasing trust in people and institutional trust, which, in turn, improves subjective mental well-being. Moreover, institutional trust is the major component of the indirect effect in both age groups, even if it has a greater effect on those over 34. Specifically, its indirect effect is equal to 34% of the total effect for young people and 49% for the others, with a differential of 14% points. Adults over the age of 34 benefit more from the effect of institutional trust on mental well-being, induced by positive perceptions of the QGS. Trust in People is another dimension of trust that significantly mediates the effect of QGS, by amplifying its positive impact on WHO-5. In contrast with Institutional Trust, this type of trust accounts for a major indirect effect for young people, mediating 18% of the total effect, whilst in the over 34 age group, it mediates only 10% of the overall effect of government support.

Discussion
This study goes beyond the usual link between good governance and economic outcomes, confirming the validity of previous research examining how the quality of institutions affects health and subjective well-being, especially in a time of emergency such as the pandemic. Positive evaluations of the perceived QGS, relating to the responsiveness and reliability of state capabilities, integrity and equity values, such as transparent and fair rules for access to income support measures, significantly increase both trust and subjective well-being (WHO-5).
Our findings also cast light on the fact that most of the relationship between QGS and WHO-5 is mainly due to the buffer effect of trust: positive evaluations of QGS promote trust in people and in institutions which, in turn, increases subjective WHO-5. This is evidence that the role of trust is fundamental for increasing levels of subjective well-being in periods of crisis, in accordance with previous empirical research ( Barrafrem et al., 2021;Roccato et al., 2021).
Finally, we find that the 18-34 age group is more influenced by the trust in other people, which accounts for a greater share of the indirect effect of QGS on WHO-5 with respect to the over 34 age group.
Although this study contributes to the literature with these novel findings, it has some limitations. The paper neglects the positive correlation between trust in institutions and trust in others: on the one hand, general trust in people can support the effectiveness of institutions and, on the other hand, institutions can condition the emergence, persistence and efficiency of trust in people, improving the situation of those in the network, without worsening that of others (Barca, 2001).
Another limitation concerns the estimated coefficients that are interpreted as average effects and should, therefore, be applied with caution to specific countries. In fact, an increasing stream of literature on the reasons of compliance to COVID-19 containment and mitigation measures highlights that there are numerous variables that may differ among countries: Maloney and Taskin (2020) indicate that, for the United States of America, much of the decrease in mobility is due to greater awareness of risk, driven by data on contagion risk; Kantor and Kantor (2020), through a cross-sectional age, sex and race stratified survey of the general USA population, point out that NPI adherence is associated with expectations and believes; in Italy the drivers of mobility responses to mitigation efforts during different restriction schemes may be the local structure of the labour market, together with other demographic factors (Gauvin et al., 2021).

Conclusion and implications
This paper highlights the importance of good governance, especially in times of crisis, for increasing trust in people and in institutions and for promoting individual well-being.
We suggest theoretical and practical implications. From a theoretical point of view, the analysis expands the list of relationships mediated by the buffer role of trust both in institutions and in other people (Lee, 2022;Chan et al., 2021;Ciziceno and Travaglino, 2019), underlining how this factor acts in the macro-micro dynamics of public intervention and individual well-being.
In terms of practical contribution, this study provides indications on the mechanism of functioning and transmission of public support. This highlights that governments should pay greater attention to integrity, fairness, reliability and responsiveness dimensions of welfare programmes. As they result in significantly increasing both trust in institutions and trust in other people, with important consequences on social cohesion, sense of community and positive mental health. The dimensions listed above are fundamental both in the design and in the governance of welfare measures.

IJSE
Possible tools useful to enhance the integrity, fairness, reliability and responsiveness values in welfare programmes should be long-term sustainability of public policy, continuity of the intervention, a clear and credible communication and a territorial capillarity of the service. That reinforces the perception of reliability and competence of policy makers and, in turn, a general sense of trust in the community.
Moreover, governments should design policies suitable for the needs of diverse age groups. The distance between institutions and younger age groups should be shortened through appropriate communication channels and social programmes in order to encourage civic participation. This is of particular interest considering that growing evidence suggests that trust attitudes, like other cultural traits, persist for long periods of time (Bjornskov, 2007).
For further development, it should be helpful to promote surveys especially for the most vulnerable groups. Granular data can help to reveal important clusters of distrust, hence, in turn, better policy responses. Surveys of this kind can also foster greater perceptions of social inclusion with respect to the most marginalised groups. Figure S1.   Table S4.
Correlation matrix between variables adopted in the moderated mediation IJSE About the authors Erica Poma is a PhD student in the doctoral research programme in Labour, Development and Innovation at the University of Modena and Reggio Emilia. Her main research interest includes gender diversity policies, trust in institutions, democracy, good governance and social policies that promote equality and individuals' well-being. Her current research focusses on the pandemic and its implications on health outcomes, with a particular attention to the impact of governments' social support. Erica Poma is the corresponding author and can be contacted at: erica.poma@unimore.it Barbara Pistoresi is Associate Professor in Economics and Head of the Master Degree in Economics, Public Policies and Sustainability of the University of Modena and Reggio Emilia (Italy). Her current research focusses on pandemic and economic crisis, subjective well-being; digital, agile, flexible work, quality of work and work-life balance.
Chiara Giovinazzo is a PhD student at the Department of Economics "M. Biagi" of the University of Modena and Reggio Emilia (Italy). Her research fields are labour economics, human capital, social mobility and skill and education mismatch in the Italian labour market. She is currently working on the impact evaluation of COVID-19 pandemic on trust and subjective well-being.
For instructions on how to order reprints of this article, please visit our website: www.emeraldgrouppublishing.com/licensing/reprints.htm Or contact us for further details: permissions@emeraldinsight.com Mental well-being and government support