1 Introduction

In general, it is argued that public policy ought to aim at improving well-being for as many people as possible since everyone ultimately wants to be happy (Frey and Stutzer 2002; Tavits 2008). For so many decades, scholarly work on what factors determine subjective well-being (SWB)Footnote 1 remained the preserve of psychologists and sociologists (Frey and Stutzer 2002; Wang and Van der Weele 2011). However, since Easterlin (1974), there has been increased research among economists on what makes people happy. In what became known as the “Easterlin Paradox,” Easterlin (1974) discovered that within and across countries, income and happiness are positively associated, but that this relationship does not hold over time (Easterlin 1974; Easterlin et al. 2010). Easterlin et al. (2010) put the Easterlin Paradox succinctly as follows:

At a point in time both among and within nations, happiness varies directly with income, but over time, happiness does not increase when a country’s income increases. We are talking here about the time series relationship of happiness and income in the long term, usually at least 10 years, sometimes more.

The relationship between income and happiness across countries has been substantially debated (see e.g., Hagerty and Veenhoven 2003; Easterlin 2005a; Hagerty and Veenhoven 2006). At the same time, the examination of non-income sources of happiness has burgeoned (e.g., Stack and Eshleman 1998; Helliwell 2001, 2003, 2006; Bjørnskov 2003a, 2008; Tavits 2008; James 2011; Hudson 2006; Leung et al. 2011; Rodriguez-Pose and von Berlepsch 2014; Sarracino 2013; Ram 2010).

Among the determinants of happiness, scholars have established relationships between social capital (as mostly proxied by trust) and happiness (e.g., Algan and Cahuc 2013; Bjørnskov 2008; Leung et al. 2011; Rodriguez-Pose and von Berlepsch 2014; Sarracino 2013; Ram 2010; Bartolini et al. 2013). For instance, Algan and Cahuc (2013) showed that across countries, generalized trust is positively correlated with many economic and social variables: per capita income, private credit (as percentage of GDP), total factor productivity (TFP), research and development, quality of legal system, and life satisfaction. Calculating country averages for trust based on the question “Generally speaking, would you say that most people can be trusted or that you need to be very careful in dealing with people?” the authors indicated that the poorest countries have the least generalized trust. Their study showed that the country averages for generalized trust for Ghana, Rwanda, and Zimbabwe are 8.5, 4.9, and 11.2 %, respectively, whereas trust is highest among Scandinavian countries of Denmark (64.5 %), Sweden (64.8 %) and Norway (68.1 %). The figure is 41.2 % in the US, 36.9 % in Great Britain, and 34.6 % in Germany.

Most studies examining the relationship between social capital and happiness focus on developed countries (see e.g., Hudson 2006; Bjørnskov 2008; Leung et al. 2011; Rodriguez-Pose and von Berlepsch 2014). Studies on the nature of this relationship among Sub-Saharan African countries are lacking. As a result, this paper investigates the relationship between social capital and SWB by examining data from the 2012 Afrobarometer Surveys (ABS), focusing on Ghana. The study examines how three measures of trust—generalized trust, interpersonal trust and institutional trust—as well as indicators of institutional quality and membership in voluntary and religious associations are correlated with absolute and relative SWB. The results reveal that there is no significant association between generalized trust and either measure of SWB. However, interpersonal trust and institutional trust are significantly correlated with both measures of SWB. Respondents who indicated being satisfied with the way democracy works in the country reported higher SWB levels. Finally, as measures of civic engagement, active membership in voluntary organizations exert a negative effect on absolute SWB but a positive effect on relative SWB, while active membership in religious associations generally diminishes relative SWB.

2 Background Literature

2.1 Social Capital and Subjective Well-Being

Coleman (1988) is often cited as one of the earliest scholars to study social capital. According to him, there are three forms of social capital (i.e., obligations and expectations, information channels, and social norms). Within this framework, the author investigated the effect of social capital on high school dropout rates, and found that these rates were low among Catholic schools, high among private schools, and that students who frequented at religious activities tended to have low dropout rates. Putnam (2000) suggested that, “social capital refers to features of social organizations such as networks, norms, and social trust that facilitate coordination and cooperation for mutual benefit.”

Subjective well-being, also a multi-faceted construct, “includes people’s emotional responses, domain satisfactions, and global judgments of life satisfaction” (Diener et al. 1999). SWB has two main components—affective and cognitive domains (Diener et al. 1999). According to the authors, the affective component of SWB has pleasant and unpleasant domains. Emotions and moods like happiness, joy, affection, etc. are in the pleasant affective domain while sadness, depression, stress, etc. belong in the unpleasant affective domain. The cognitive domain captures life satisfaction, encompassing satisfaction with past, current, and future life as well as desire to change life (Diener et al. 1999, p. 277). Helliwell and Putnam (2004) explained that even though “happiness” and “life satisfaction” measures tend to provide consistent results, “happiness” measures short-term reactions of moods and emotions, while “life satisfaction” tends to measure long-term self-evaluations that are comparatively more stable.

Several scholars have used Coleman’s ideas to examine the role of social capital on SWB. For instance, Leung et al. (2011) used Coleman’s (1988) social capital measures to create four blocks of social capital variables with which they examined the effect of social capital on happiness. Using data from the Canadian General Social Survey of Social Engagement Cycle 17 (2003), they found evidence of a positive association between social capital and happiness. Also, Rodriguez-Pose and von Berlepsch (2014) used data for about 48,583 respondents across 25 European countries to examine the effect of social capital as established by Coleman (1988) on happiness. They reported that, indeed, social capital is an important determinant of happiness; and that most important among these determinants were informal social interactions, as well as generalized and institutional trust.

Research on social capital has also frequently cited Putnam’s (2000) work (e.g. Glaeser et al. 2002; Bjørnskov 2006, 2008; Rodriguez-Pose and von Berlepsch 2014). Lamenting the decline in civic engagement in America, Putnam (2000) argued that, analogous to physical and human capital, social capital through civic engagement and social interconnectedness affects economic development, reduces social crime and improves effectiveness of governance. Other studies provide further evidence on the positive role of social capital. For instance, Helliwell (2001) reported a significant association between social capital and economic growth, health outcomes, and SWB. If trust triggers economic growth (see Zak and Knack 2001), then trust invariably leads to happiness (Bjørnskov 2003a).

According to Bjørnskov (2003a), cross-country evidence suggests that five Nordic countries, Switzerland, and the Netherlands are among the few happiest countries in the world. Examining the determinants of happiness, including economic possibilities, inequality, uncertainty, democratization, and social capital, the author found evidence to support the study’s hypothesis that “social capital has an independent effect on people’s happiness” (p. 3). In particular, he argued that the abundance of social capital in Switzerland and Nordic countries accounted for their distinctly higher levels of happiness relative to the rest of the world. In another study, Bjørnskov (2008) used US Census data for the period 1983–1998 to investigate the relationship between social capital and happiness and found that social trust directly affects happiness. However, he noted that only when one examines the impact of informal sociability (e.g., going swimming, going to movies, entertaining people at home, etc.) at the regional level is there a positive relationship.

Estimating the importance of social capital for happiness, Helliwell (2006) uncovered that “respondents who have frequent contacts with family, friends and neighbors have SWB almost a full point higher, on the 10-point SWB scale, than those with no such contacts.” Portela et al. (2013) provide further evidence on the social capital-happiness relationship. They used data from the 4th wave of the European Social Survey for empirical analysis and observed that social networks, social trust, and institutional trust were significantly correlated with SWB. In a cross-country survey, Ram (2010) used various model specifications and variable proxies to examine the effect of social capital on happiness. The study’s results were inconclusive. Of about 10 models estimated, trust was significantly associated with happiness in one equation (p < 0.10), indicating a weak effect. This led the author to conclude that the effect of social capital (as measured by trust) on happiness is “extremely fragile” (p. 414). Among other recommendations, Ram suggested that alternative measures of social capital should be explored.

While certain measures of social capital are often generally found to have a positive and significant effect on happiness, cross-country comparisons, especially between rich and poor societies or countries sometimes reveal inconsistent results. In other words, the same social capital measures might have significant effect on happiness in rich countries but no effect in poor countries. An example is provided by Sarracino (2013) who focused on social capital variables such as spending time with relatives, friends from work, people from church and sports and involvement in voluntary organizations including religious organizations, arts, sports, political, charity, as well as professional organizations. Sarracino observed that whereas spending time with colleagues from work, people from religious environment and participation in political or professional organizations are correlated with happiness among respondents in low-income countries, this correlation held only for spending time with people from religious environment, recreational/sport environments and with colleagues from work among high-income country respondents. Furthermore, the results showed that trust is significantly correlated with happiness in high-income countries but not in low-income countries.

Examining the predictors of SWB in Ghana, Addai et al. (2013a) used data from the World Values Survey (WVS) in their study. Using personal happiness and life satisfaction as measures of SWB, Addai and his colleagues found no evidence of association between interpersonal trust, institutional trust and civic involvement on the one hand and SWB on the other hand. Their results, however, revealed that community engagement—measured by activism, reading news from various sources, and voting in previous elections—is positively and significantly correlated with both personal happiness and life satisfaction.

2.2 Summary and Research Questions

The literature on the relationship between social capital and SWB mostly indicates that social capital enhances well-being. A few studies have, however, found that trust does not exert a significant effect on SWB. Furthermore, differences exist between high- and low-income countries in terms of what components of social capital contribute to well-being (Sarracino 2013). Because most studies are focused on developed countries, little is known about how social capital affects well-being in less developed countries, especially Sub-Saharan Africa. By conducting an empirical investigation of the relationship between social capital and SWB in Ghana using data from the 2012 ABS, this study provides an African perspective on the social capital—SWB relationship.

The general research question this paper seeks to address is whether social capital and SWB are correlated in Ghana. Because social capital is a multi-faceted construct, a number of social capital measures are examined. First, because most studies in the literature suggest that trust is positively and significantly correlated with SWB, various forms of trust and how they are correlated with SWB are explored. Second, how measures of institutional strength and civic engagement affect SWB of Ghanaians is also examined. Therefore, the specific research questions are as follows: Do people who trust other people report significantly higher SWB levels? Do people who trust public institutions report significantly higher SWB levels? As measures of institutional quality, how are perceived corruption, experienced corruption, and satisfaction with democracy correlated with SWB? Finally, does civic engagement influence well-bing?

3 Ghana: Economy, People and Culture

Formerly The Gold Coast under British Colonial rule, modern day Ghana is a West African nation located along the Gulf of Guinea lying between 6° and 11°N and 1°E and 3°W (see e.g., Breuning-Madsen and Awadzi 2005). It covers a total land area of 92,099 square miles (238,535 km2) which is about the size of the United Kingdom (Robson and Freel 2008), sharing borders with Ivory Coast to the west, Togo to the east and Burkina Faso to the north. According to the 2010 Population and Housing Census, Ghana’s total population in 2010 was 24,658,823 (Ghana Statistical Service 2012). Its people are diverse in many respects. In terms of ethnicity, there are many ethnic groups, with the Akans constituting about 47.5 % of the total population, Mole Dagbani (16.6 %), Ewe (13.9 %) and Ga-Dangme (7.4 %) (Ghana Statistical Service 2012). In 2010, only 4.7 % of Ghanaians were 65 years and older, while about 38.5 % were below 15 suggesting that the country is very youthful. Regarding religion, 71.2 % of Ghanaians are Christian, 17.6 % are Muslim, and 5.2 % are traditional believers while 5.3 % have no religious affiliation (Ghana Statistical Service 2012). Religion is believed to be very important in all aspects of life in Ghana (e.g., Mbiti 1969; Pokimica et al. 2012; Addai et al. 2013b). Arguably, the lack of social welfare programs in Ghana makes religion and religious participation a source of social support that people resort to in times of life’s difficulties (Pokimica et al. 2012).

Since gaining independence in 1957 (De Smith 1957), the country has undergone substantial political and economic transformations. It experienced its first coup d’état in 1966 that brought the civilian rule, following independence, to an end. Military and civilian regimes prevailed intermittently between February 1966 and January 1993 (Anaman and Agyei-Sasu 2012). Since 1993, Ghana has enjoyed stable democratic processes, having successfully undergone six (6) successive peaceful Presidential and Parliamentary elections. The two dominant political parties are the National Democratic Congress (NDC)—the incumbent—and the New Patriotic Party (NPP), the main opposition party.

With a 2012 GDP per capita income of about $1,600 (World Bank 2013), Ghana is now a lower middle-income country. The country had higher GDP per capita than South Korea in the 1960s (Robson and Freel 2008) but now trails behind South Korea in GDP per capita. South Korea’s GDP per capita in 2012 was $22,590.16 (World Bank 2013) which is about 14 times that of Ghana. Political crises as well as macroeconomic shocks in the 1960s and 1970s caused substantial economic decay that the country had to adopt the Economic Recovery Programme at the directives of the World Bank and the International Monetary Fund in 1983, and subsequently the Structural Adjustment Programme (Aryeetey et al. 2000). Although Ghana started producing and exporting oil in December 2010 (see e.g., Ackah-Baidoo 2013), agriculture remains the mainstay of the economy, serving as the primary source of livelihood for about 45.8 % of Ghanaians, with 95.1 % of them involved in smallholder crop farming (Ghana Statistical Service 2012).

4 Conceptual Framework

The benefits of social capital are many (e.g., Helliwell and Putnam 2004; Zak and Knack 2001). For instance, the health benefits of social capital have been studied (e.g., Kawachi 1999; Poortinga 2006). Kawachi (1999) noted that social capital could have positive outcomes for both individual and community health. Poortinga (2006) found that social capital significantly boosts both self-reported and actual health. Indeed, health is among Wilson’s (1967) characteristics of a typically happy person. Because health turns out to be one of the most important determinants of SWB (Gerdtham and Johannesson 2001), by improving health outcomes (both perceived and actual), social capital positively influences well-being.

Fear of crime and crime victimization have been shown to significantly reduce well-being (e.g., Denkers and Winkel 1998; Kuroki 2013; Powdthavee 2005; Møller 2005; Davies and Hinks 2010; Sulemana 2014). Individuals who live in fear, say in their neighborhood, may feel that they have lost control over their life. Because control over life significantly enhances well-being (Verme 2009), fear of crime and crime victimization negatively impact SWB. Also, feeling unsafe may increase psychological stress, reduce physical activity, and increase obesity because people would tend to stay indoors (Burdette et al. 2006). However, in neighborhoods where social capital is dense, chances of crime are lowered, consequently reducing fear of crime or feelings of unsafety (Helliwell and Putnam 2004). Helliwell and Putnam (2004) contend that the existence of barbecues or neighborhood associations can deter crime, thereby benefiting a larger community, members of whom may not necessarily be directly involved in these social networks or activities.

Ties with family, neighbors, and friends also have benefits for the individual. The unpleasant domain of the affective component of SWB takes into account stress, depression, sadness, loneliness, etc. (Diener et al. 1999). But as Helliwell and Putnam (2004) put it, “People who have close friends and confidants, friendly neighbors and supportive coworkers are less likely to experience sadness, loneliness, low self-esteem and problems with eating and sleeping.” Thus, social capital would tend to exert positive influences on SWB. Social capital has also been linked to educational performance, lower public corruption, and efficient governance, among others, which scholars have found to positively affect well-being (e.g., Putnam 2000; Bjørnskov 2003b). Finally, Zak and Knack (2001) demonstrated that by reducing transaction costs, trust increases investment and consequently leads to economic growth. They argued that low-trust societies tend to grow much less than high-trust societies. Because income positively affects well-being (Easterlin 1974; Easterlin et al. 2010), by increasing output, trust could increase well-being too.

5 Data and Empirical Strategy

To empirically test the relationship between social capital and SWB in Ghana, this study uses data from the Round 5 of the ABS (2012) in which 2,400 respondents were interviewed in Ghana. Jointly sponsored by The Institute for Democracy in South Africa (IDASA), Ghana Center for Democratic Development (CDD-Ghana) and Michigan State University (MSU), these surveys were conducted in many countries in Africa. In Ghana, like other participating countries, respondents were randomly selected to constitute a representative sample of the adult population (18 years and older). They were then interviewed on a wide array of issues pertaining to respondents’ views about democracy, corruption, how the economy is managed, among others. The interviews were face-to-face and conducted in five languages—English, Akan, Ga, Ewe and Dagbani. The overall response rate was 72.9 %.

Scholars have argued that the ABS are reliable and have validity (see e.g., Bratton 2007; Razafindrakoto and Roubaud 2010; Mattes 2008). Investigating the validity of ABS, Bratton (2007) compared assessment of the surveys’ “extent of democracy” with the Freedom House Index, and found that the two research methods are consistent with a Pearson’s correlation coefficient of higher than 0.8. On perceptions of corruption in Africa by household surveys and expert opinions, Razafindrakoto and Roubaud (2010) found a positive correlation between ABS and the Global Corruption Barometer results. Based on factor and reliability analyses, Mattes (2008) reported that responses in ABS have high degree of face validity, content validity, construct validity as well as reliability (internal consistency).

5.1 Dependent Variables

Ideally, the SWB measure to use should be one identical to the WVS question “Taking all things together, would you say you are: very happy, quite happy, not very happy, not at all happy” (see e.g., Schyns 1998) or “Taking all things together, how satisfied are you with your life as a whole these days? Are you very satisfied, satisfied, not very satisfied, not at all satisfied?” (e.g., Kahneman and Krueger 2006). Instead, the ABS question asks respondents about their “living conditions.” Specifically, respondents were presented with the following: “Let’s discuss economic conditions. In general, how would you describe: Your own present living conditions?” The responses include “Very good”, “Fairly good”, “Neither good nor bad”, “Fairly Bad”, “Very Bad”, and “Don’t know.” Clearly, this question asks respondents about the subjective evaluation of their own present circumstances. SWB pertains to “people’s cognitive and affective evaluations of their lives” (Diener 2000). Different cultures may interpret SWB questions differently, especially questions about “happiness” (e.g., Uchida et al. 2004). Hence, although SWB may be universal across cultures to some extent, it is important to also understand SWB in culture specific contexts (Tov and Diener 2009). In a developing country with many different dialects such as Ghana, it could be argued that asking people about their living conditions taps into the way they evaluate their lives. For instance, a similar survey question measuring life satisfaction that appeared in surveys in South Africa asked respondents “Taking everything into account, how satisfied is this household with the way it lives these days? 1 = Very dissatisfied, 2 = dissatisfied, 3 = neutral, 4 = satisfied, 5 = very satisfied” (see e.g., Bookwalter et al. 2011). Therefore, following the precedent of other studies (Pokimica et al. 2012; Sulemana 2014), I treat responses to this question as indicative of absolute SWB.

Respondents were also asked about their living conditions relative to the living conditions of other Ghanaians: “In general, how do you rate your living conditions compared to those of other Ghanaians?” Responses were “Much worse”, “Worse”, “Same”, “Better”, “Much better”, and “Don’t know.” Because this examines how respondents perceive their living conditions compared to those of other Ghanaians, and again, following the precedent of Pokimica et al. (2012), I treat this as relative SWB. Both absolute SWB and relative SWB are ordinal variables.

5.2 Explanatory Variables

As explanatory variables, social capital measures are constructed in several ways.

(1) Generalized trust—This is constructed based on responses to the question “Generally speaking, would you say that most people can be trusted or that you must be very careful in dealing with people?” The use of responses to this question as a measure of generalized trust seems to be commonly accepted in the literature and as a proxy for social capital (e.g. Zak and Knack 2001; Bjørnskov 2006; Ram 2010; Portela et al. 2013). As argued by Portela et al. (2013), the majority of studies on social capital have focused on generalized trust. (2) Interpersonal trust—This variable is based on responses to the question: “How much do you trust each of the following types of people? Your relatives; your neighbors; other people you know. 0 = Not at all, 1 = Just a little, 2 = Somewhat, 3 = A lot.” A summed–score procedure was used to construct interpersonal trust as a composite index of trust in people (see e.g., Addai et al. 2013b). This is also referred to as dyadic trust—“trust in individuals or specifically identified individuals” (Hudson 2006). (3) Institutional trust—Following the convention in the literature (e.g., Hudson 2006; Portela et al. 2013; Addai et al. 2013b), this variable measures how much trust people have in public institutions. Specifically, respondents were asked to indicate how much they trust The President and the Office of the President, Parliament, Electoral Commission (EC), the Army and Law Courts (0 = Not at all, 1 = Just a little, 2 = Somewhat, 3 = A lot). A composite index of trust in institutions was then created. These three measures of social capital—generalized, interpersonal, and institutional trust—fall under the umbrella term “social trust” (Portela et al. 2013, p. 495).

(4) Perceived corruption, experienced corruption, and satisfaction with democracy—Bjørnskov et al. (2010) demonstrated that institutional strength (as measured by variables such as honesty and efficiency of government, democratic process, legal quality, etc.) could also be taken into account when using trust as a proxy for social capital.Footnote 2 How clean a government is could be determined in part by the levels of perceived and experienced corruption. Also, how satisfied people are with democratic processes may affect their trust in their government. Therefore, corruption perceptions and experiences, as well as the respondent’s satisfaction with how democracy works in the country are included. (5) Membership in voluntary and religious associations—To account for social interconnectedness and civic engagement (Putnam 2000; Portela et al. 2013), these variables were included to capture whether the respondent is an active member or official leader of a voluntary or religious association. Indeed, Putnam (2000) states, “Trust and engagement are two facets of the same underlying factor—social capital.”

5.3 Control Variables

The following variables were included as control variables: (a) Individual demographic characteristics such as age, gender, education, as well as their employment status (see e.g., Clark et al. 1996; Di Tella et al. 2001). (b) Respondent’s political affiliation—Individuals are happy when their preferred political party is in power (Tavits 2008). (c) Religiosity and religious denomination—People who believe that religion is important in their life and participate in religious activities tend to report higher levels of SWB (e.g., Helliwell 2003; Pokimica et al. 2012; James 2011). (d) Ethnicity/tribal group—It has been reported that blacks and non-whites are significantly less happy than white Americans (e.g., Blanchflower and Oswald 2004). In the case of Ghana, tribal differences are reported to be associated with SWB (e.g., Pokimica et al. 2012; Addai et al. 2013a). (e) Region of residence—Due to cultural and economic differences, a respondent’s region of residence might matter for their well-being (Pokimica et al. 2012; Addai et al. 2013a). (f) City size—This indicates whether the respondent lives in an urban versus a rural area. Because urban areas generally have more social amenities in Ghana (particularly education and health), urban residents might be happier (Sulemana 2014). (g) Confidence in the economy– This indicates optimism as a personality trait which enhances well-being (e.g., Scheier and Carver 1992; Helliwell 2006). (h) Perceived neighborhood insecurity—If people feel unsafe in their neighborhood, they are likely to remain indoors, which may increase psychological stress, and reduce physical activity, and cause obesity (Burdette et al. 2006) and thereby reduce well-being (Lehman 1983).

Among other significant correlates of SWB often cited in the literature are marital status (e.g., Lee et al. 1991; Stack and Eshleman 1998; James 2011), children (Glenn and Weaver 1979; Argyle 1999), satisfaction with household finances (e.g., Stack and Eshleman 1998; James 2011) and health status (Stack and Eshleman 1998; James 2011). However, due to data limitations (unavailability), these variables were not examined here.

The empirical strategy follows the precedent of scholars studying the relationship between SWB and its correlates (e.g. Helliwell 2003; Bjørnskov 2003a; 2006; 2008; Portela et al. 2013; Ram 2010; James 2011) by treating SWB as the dependent variable and social capital as explanatory variable while controlling for demographic characteristics and other significant correlates of happiness. Therefore, the formal econometric model to be estimated is:

$${\text{SWB}} =\upalpha + \text{X}^{\prime }\upbeta + Z^{\prime }\upphi +\upvarepsilon$$

where SWB denotes SWB of the respondent (absolute and relative), X is a vector of social capital measures, Z is a vector of control variables, α, β, and ϕ are vectors of parameters to be estimated, and ε is the error term. Since each of the dependent variables is ordinal in nature, the model is estimated using ordered logit (see e.g., Ferrer-i-Carbonell and Frijters 2004; Portela et al. 2013). Ferrer-i-Carbonell and Frijters (2004) have argued that assuming cardinality or ordinality of well-being measures makes little or no difference. Furthermore, because OLS estimates are easier to interpret relative to ordered logit/probit results (e.g., Kuroki 2013), the equation is also estimated using OLS.

6 Empirical Results

Table 1 presents the variables used in the study along with their descriptions and summary statistics. On a 5-point scale, the average absolute SWB score is 2.580, which is below the neutral score of 3 (Neither good nor bad). On the other hand, the relative SWB mean score of 3.02 indicates that the average respondent rated their living conditions compared to other Ghanaians as “same.” Figures 1 and 2 show the distributions of responses to the SWB questions. In Fig. 1, about 55.6 % of respondents rated their living conditions as either “very bad” (30.464 %) or “fairly bad” (25.143 %). Those indicating “Fairly good” and “Very good” constitute 38.03 % while about 6.36 % indicated “Neither good nor bad.” In Fig. 2, about 38.76 % of them rated their living conditions as either “Better” or “Much better” than those of other Ghanaians while 32.4 % think their conditions are “Worse” or “Much worse.”

Table 1 Variable descriptions and summary statistics
Full size table
Fig. 1
figure 1

Distribution of responses to the question. “In general, how would you describe: Your present living conditions?”

Full size image
Fig. 2
figure 2

Distribution of responses to the question. “In general, how do you rate your living conditions compared to those of other Ghanaians?”

Full size image

With regards to the social capital measures, only 10.9 % of respondents think most people can be trusted. The average scores for interpersonal and institutional trust are 5.214 and 10.174, respectively. The mean score of 7.285 out of 8 for corruption perceptions suggests that about 91 % of respondents think that corruption is widespread among public officials in Ghana. This contrasts sharply with about 8.5 % for experienced corruption.Footnote 3 Also, respondents seem to be satisfied with the way democracy works in the country. About 20 % of respondents are active members/leaders of a voluntary association while 41.4 % of respondents belong to a religious association.

Table 2 reports the OLS and ordered logit regression results when each SWB measure is regressed on measures of social capital, while controlling for other variables. As the table shows, the results from both estimation techniques are identical. Generalized trust is not significantly associated with either measure of SWB. While generalized trust is often found to significantly enhance SWB, the finding that it does not exert a significant effect on SWB is consistent with results from some previous studies (e.g., Ram 2010; Sarracino 2013). Interpersonal trust and institutional trust are however significantly correlated with SWB (p < 0.01). Both OLS and ordered logit regression results indicate that interpersonal and institutional trust positively and significantly affect absolute SWB (p < 0.01). On the other hand, while interpersonal trust exerts a positive and significant effect on relative SWB (p < 0.05) in the ordered logit model, institutional trust is significant (p < 0.05) for both models.Footnote 4 Evidence in the literature corroborates these findings. People who are more trusting of others and their institutions are happier (see Hudson 2006; Helliwell and Huang 2008).

Table 2 OLS and ordered logistic regression results for the effect of social capital on absolute and relative subjective well-being
Full size table

Neither perceptions about corruption among public officials nor experiences of corruption (having paid a bribe, given a gift, or done a favor for a government official) is significantly correlated with absolute SWB. We would expect that bribe victimization would have a negative effect on SWB (see e.g., Gillanders 2011; Graham 2011; Singer 2013). An explanation offered by Graham (2011) is that in high corruption societies where bribery is how people “get things done”, bribery does not reduce well-being. The results show that perceived corruption is positively associated with relative SWB (p < 0.05), suggesting that people who report that public corruption is widespread also rated their living conditions as better than those of other Ghanaians. The evidence in the literature suggests that democracy and quality of political institutions affect happiness directly (Frey and Stutzer 2000; Samanni and Holmberg 2010). The results here confirm this as respondents who indicated being satisfied with the way democracy works in Ghana also reported higher absolute and relative SWB (p < 0.01). While membership in voluntary associations is negatively associated with absolute SWB (p < 0.01), it is positively associated with relative SWB (p < 0.05). Also, respondents who reported being a member of a religious association reported lower levels of well-being compared to other Ghanaians (p < 0.01), even though such membership is not significantly correlated with absolute SWB.

Regarding the control variables, consistent with previous findings, I find that age has a negative, significant (p < 0.01) and U-shape relationship with absolute SWB (see e.g., Frijters and Beatton 2012; Blanchflower and Oswald 2008). However, age has no relationship with relative SWB. There are no significant gender differences in the well-being of Ghanaians. Consistent with other studies (e.g., Wilson 1967; Gerdtham and Johannesson 2001; Cuñado and de Gracia 2012), the results also indicate that, generally, respondents with more education are happier (p < 0.01). It is established that unemployment diminishes happiness (e.g., Stack and Eshleman 1998; Frey and Stutzer 2000; Helliwell 2003; Krueger and Mueller 2010; James 2011; Ohtake 2012). The results here confirm that unemployment reduces absolute SWB (p < 0.01), although it has no significant effect on relative SWB.

With respect to the effect of political representation on SWB, the study notes that, relative to other political parties, respondents who are affiliated with the NPP reported significantly lower levels of absolute SWB (p < 0.01) and relative SWB (p < 0.05). Those affiliated with the NDC did not report higher SWB. Using data for about 68 European countries, Tavits (2008) found that individuals report higher levels of SWB when their choice of political party is in power. However, she noted that if the government is believed to be corrupt, then having one’s party of choice in power does not increase well-being. Thus, a possible explanation for the result on the NDC coefficient is that these respondents perceive corruption as pervasive in the government. Also, Frey and Stutzer (2000) contend that the positive effect of democracy could be explained by political outcomes that come closer to the voters’ preferences. By extension, therefore, when political outcomes do not align with voters’ preferences, those individuals are likely to be unhappy; hence the negative and significant association between affiliation with the NPP and SWB.

Religiosity, a measure of how important religion is in one’s life, is typically found to have a positive impact on SWB (e.g. Stack and Eshleman 1998; James 2011; Deaton and Stone 2013; Pokimica et al. 2012). However, I find no evidence of a significant association between religiosity and SWB. In addition, there are no significant differences among the various religious groupings as far as absolute SWB is concerned, although Protestants reported lower relative SWB levels (p < 0.01). Pokimica et al. (2012) found that both religiosity (religious importance) and religious denomination (Catholic, Protestant, Muslim, etc.) have “weak-to-moderate” significant association with both absolute and relative SWB. Explaining these results, the authors contend that, in Ghana, the lack of social support programs means that people resort to religion (religiosity and religious organizations) as insurance during economic and social distress (p. 63). Furthermore, Deaton and Stone (2013) asserted that “Religious Americans are happier, and religious states in America are happier.” Yet the results in this study do not suggest a systematic, positive, and significant relationship between religiosity and SWB. Similar to the findings here, Addai et al. (2013a) reported an insignificant effect of religiosity on SWB. They explain that this might be due to the fact that most of the respondents reported that they are “very religious.” As indicated on Table 1, the average score for religiosity was 3.867 on a 4-point scale, suggesting that over 90 % of respondents indicated that religion is important in their life.

Relative to other ethnic groups, the results reveal that Akans are the least happy about their living conditions (p < 0.01 for the OLS model and p < 0.05 for the ordered logit model). This finding contradicts earlier results (e.g., Pokimica et al. 2012; Addai and Pokimica 2010; Addai et al. 2013a). These studies all arrived at the conclusion that Akans were significantly happier than all other ethnic groups in Ghana. As stated above, the Akans are the largest ethnic group in Ghana. The NPP is noted as a predominantly Akan party (see e.g., Nugent 1999). Since political representation affects SWB, and since the NPP was in opposition at the time of the survey, this could explain why the Akans were the least happy ethnic group.Footnote 5 Compared to residents of the Volta Region, residents of all other regions, except Greater Accra Region, reported significantly lower levels of absolute SWB (p < 0.01). Residents in all regions indicated lower levels of relative SWB compared to the residents of the Volta Region (p < 0.01). As expected, rural–urban differences affect SWB in Ghana. Urban dwellers reported better living conditions than residents of rural areas (p < 0.01). This might be due to the fact that urban areas tend to enjoy more social amenities than the rural areas. The results also reveal that people who are optimistic about the economy indicated being happier (p < 0.01), while feeling unsafe in one’s neighborhood reduces SWB (p < 0.01 for the OLS model and p < 0.05 for the ordered logit model).

7 Discussion

This paper sought to examine the relationship between social capital and SWB in Ghana. Several measures of social capital were used in this exercise. The results suggest a lack of systematic relationship between generalized trust and SWB in the Ghanaian context. Ram (2010) stated, “In most cases, the parameter for generalized trust, which is used as a proxy for social capital, lacks statistical significance at the conventional levels” (p.410). Halpern (2005) argued that high-trust nations are probably happier because they are wealthier, and not because they are more trusting. This argument was based on the idea that once wealth is controlled for, the effect of trust on happiness diminishes. Indeed, Ghana is a low-trust society. The literature suggests an important effect of trust on numerous socio-economic phenomena (e.g., Algan and Cahuc 2013; Zak and Knack 2001). Yet, as reported in Table 1, only 10.9 % of respondents believe that “most people can be trusted.” Using WVS data, Algan and Cahuc (2013) estimated this proportion to be 8.5 %. On the other hand, interpersonal trust is significantly correlated with well-being. Further examination of the data reveals a declining trend as we move from trust in relatives to trust in “most people.” Specifically, trust in relatives, neighbors, and other people respondents know are 74.2, 52.8, and 41.3 %, respectively, while generalized trust is at a low of 10.9 %. Banfield (1958) made the argument that people trust people in their families but do not trust others outside of their families.

To the extent that institutional trust positively affects well-being, government ought to build more trust in public institutions by increasing transparency and reducing corruption, especially corruption perceptions because corruption destroys trust (You 2005; Rothstein 2010). Perceived corruption is significantly associated with relative SWB, but not with absolute SWB, while experienced corruption has no significant relationship with either measure of SWB. It could be that because corruption perceptions are widespread in the country, people do not feel negatively affected by corruption. Some scholars have argued that African countries, including Ghana, have embraced corruption as an aspect of everyday life (e.g., Gould and Amaro-Reyes 1983). This might also explain the positive and significant effect of perceived corruption on relative SWB. In spite of the high corruption perceptions, majority of Ghanaians are satisfied with the country’s democratic processes. Thus, the positive and significant effect of satisfaction with democracy supports previous findings that democracy enhances well-being (e.g., Frey and Stutzer; Dorn et al. 2007).

Previous studies have also documented that participation in voluntary and religious organizations enhances well-being (e.g., Bruni and Stanca 2008) although other studies find no effect or inconsistent results (e.g., Yip et al. 2007; James 2011). The findings here partly support this, but remain puzzling. James (2011) obtained a negative but insignificant coefficient for the effect of voluntary organizations on well-being for Mexico, a positive but weak relationship for Canada, and positive but insignificant effect for the USA and Brazil. Thus, it seems that this effect has cross-country variations. Also, while the question about membership in a religious association was explicit and unambiguous, the one on voluntary associations was not specific. Respondents were asked to indicate whether they are an official leader, an active, inactive, or not a member of “some other voluntary association or community group” without specifying what “other” is being referenced. This makes it difficult to probe further why this variable has inconsistent and unexpected effect on SWB.

The results for membership in religious association, religious importance, and religious denomination may warrant further investigation. Pokimica et al. (2012) found positive but weak-to-moderate effect of religiosity and religious denomination on SWB while Addai et al. (2013a) obtained insignificant effect of religiosity on happiness and life satisfaction. However, Addai et al. (2013a) obtained significant effects of religious denomination and participation on these SWB measures. The results in the present study indicate that although membership in a religious group has no effect on absolute SWB, it exerts a negative effect on relative SWB. Could it be that such membership is triggered by relatively worse living conditions? For instance, some Ghanaians resort to religion whenever they face life’s challenges (e.g., Asamoah-Gyadu 2005; Pokimica et al. 2012). Asamoah-Gyadu (2005) recounted how an evangelist was invited to travel from London to Ghana to lead a prayer crusade aimed at solving the problems of the now defunct Ghana Airways. More recently, a pastor and his congregation prayed for the Ghanaian Cedi (the domestic currency) to stop depreciating against major global currencies (MyJoy Online 2014). Ruiter and Van Tubergen (2009) found that religious participation is affected by personal and societal insecurities. In that study, the authors noted that across countries, low income individuals and the unemployed are more likely to attend religious meetings, and that people with more education tend to be less frequent at religious meetings (p. 886). Therefore, it could be argued that the negative and significant relationship between membership in a religious association and relative SWB reflects the notion that they are participating because their living conditions are worse than those of other Ghanaians.

The study also reveals some underlying cultural, political, socio-economic differences among Ghanaians. One interesting finding is that being an Akan is negatively and significantly correlated with SWB. This contradicts previous evidence where the Akans were found to be the happiest (e.g., Addai and Pokimica 2010; Pokimica et al. 2012; Addai et al. 2013a). So what has changed? Addai and Pokimica (2010) and Pokimica et al. (2012) used data from the 2008 ABS while Addai et al. (2013a) used the wave 5 of the WVS conducted between 2005 and 2008. Indeed, both surveys were conducted when the NPP, a predominantly Akan party, was in power. The present study however used data from the 2012 ABS which were conducted between 2011 and 2012 when the NDC was in power. Therefore, the political representation—SWB argument appears to be at work in Ghana in that the Akans, who previously were the happiest when their preferred party was in power, now reported lower SWB levels when it went into opposition.

8 Summary and Conclusion

This paper provides an empirical investigation of the relationship between social capital and SWB in Ghana. Using data from the 2012 ABS, the paper examined how generalized trust, interpersonal trust, institutional trust, institutional quality proxies and membership in voluntary organizations affect absolute and relative SWB among Ghanaians. The results indicate that generalized trust has no significant effect on SWB in Ghana, after controlling for individual demographic characteristics and other important determinants of SWB. However, interpersonal trust and institutional trust are significantly correlated with SWB. Membership in a religious organization exerts a negative effect on relative SWB.

Since generalized trust is the commonest proxy for social capital (e.g., Portela et al. 2013; Ram 2010), and because other measures of social capital are significantly associated with SWB, these results ought to be interpreted with caution. That said, the results have policy implications. In developing countries where institutions are not very developed, especially where property rights are loosely defined, and transaction costs of using formal legal systems are very high, social capital (trust, social networks, etc.) are very essential (e.g., Lyon 2000). Moreover, because generalized trust has been shown to influence many socio-economic variables, including economic growth (Algan and Cahuc 2013; Zak and Knack 2001), it is worrisome that only about 10.9 % of Ghanaians—and 8.5 % from WVS data (see Algan and Cahuc 2013)—think that most people can be trusted in Ghana. These figures suggest that distrust among Ghanaians is very high. Therefore, public policy targeted at increasing trust among Ghanaians would be seemly.

If the goal of government is to make life better for citizens (Easterlin 2005b; Tavits 2008), then as Sarracino (2010) argues, public policy should not only focus on promoting economic growth, but should also focus on building social capital, for by increasing trust among citizens, public policy could achieve multiple outcomes: economic growth, increased happiness, among others. A plausible channel for increasing trust is through education. By incorporating the importance of trust in the educational curricula, public policy could make some significant gains in increasing trust among Ghanaians.