1 Introduction

Over the past half-century, the post-World War II “liberal international economic order” has facilitated an extraordinary worldwide spread of prosperity. It has considerably improved the material well-being of millions of people around the globe. Unfortunately, not all regions of the world have benefited from such amazing prosperity. Sub-Saharan Africa is one region that has barely participated in this great economic miracle. Despite the global prosperity, the region has been besieged by economic troubles, wars, and HIV and AIDS epidemic resulting in marked economic retrogression, decline in quality of life, and even reversals in social development. It is therefore not surprising that, in the mist of unprecedented global prosperity, Sub-Saharan Africa has remained as the poorest region of the world. All indicators point to sad reality that it also seems to be the only region that has been slipping further behind over the past decades rather than advancing in terms of material well-being.

The region’s precarious situation has been made worse by a recent global economic crisis. The crisis which hit in the wake of the food and fuel crises in 2007-08, is having a major impact on Sub-Saharan African countries through declines in commodity prices, tourism earnings, exports, remittances, and private capital flows. According to the International Monetary Fund Report (IMF) (2009) remittance inflows, which were about $20 billion a year to the region before the financial crisis, have fallen by 4–8 percent. Private capital flows—which had surged to $53 billion in 2007 and were financing much-needed infrastructure and commodity-based investments—fell by 40 percent in the second half of fiscal 2009.

The economic crisis has significantly undermined the economic growth and well-being in Sub-Saharan Africa and is likely to make things worse in the near future without some miraculous intervention. Growth in the region which had accelerated from 3.1 percent in 2000 to 6.1 percent in 2007, is now projected at only 1.7 percent for 2009—down from the projected 6.4 percent, and far below the average growth rates of 5.3 percent posted by the continent’s best 15 performing countries for more than a decade. This will slow progress toward the Millennium Development Goals (MDGs) (IMF 2009).

Despite all the challenges, some countries in the region, such as Uganda and Ghana, have been able to chalk some economic success. Key to poverty reduction, growth had accelerated in such countries as a result of improved macroeconomic policies brought about by decades of economic reforms, favorable commodity prices, and significant increases in aid, capital flows, and remittances. These countries have also supported their economic reforms and performance by improvements in governance and accountability. With the help of these countries, the continent is making headway toward reducing poverty and achieving the MDGs among pockets of the population. Thanks to these countries the proportion of Africans living on less than $1.25 a day fell from 58 percent in 1996 to 50 percent in the first quarter of 2009. The prevalence of HIV/AIDS stabilized, primary school enrollment increased, and progress was being made in other areas of human development (IMF Report 2009).

All indicators point to the reality that sustaining Sub-Saharan African countries economic progress will require staying the course of economic reforms of the past decades. The challenge facing Sub-Saharan African countries then is how to keep the momentum of reforms going and at the same time improve the well-being of the people so as to avert social and political instability which normally follow reforms in poor countries. The task is made more difficult as these countries struggle to weather the storm brought about by the worst global recession since the Great Depression. To avert instability in these countries in the mist of global crisis, because of the deterioration in the economic well-being of individuals, there is a need to understand internal economic and non-economic factors shaping day-to-day economic well-being of the people. One factor which political scientists and economists have implicated in instability, poor economic growth, and, hence, worse economic well-being of the people in Sub-Saharan Africa is ethnicity.

2 Ethnicity and Economic Growth: Review of the Literature

Studies have explored the interconnectedness between social heterogeneity and economic growth and, hence, well-being across countries and have reported an inverse relationship between them (Easterly and Levin 1997; Mauro 1995; Montalvo and Reynal-Querol 2005; Barro 1991). Although a number of conceptual ideas have been articulated to explain the negative effect of heterogeneity on economic growth, some investigators have proposed that ethnic polarization plays an important role (Easterly and Levin 1997; Barro 1991; Tavares and Wacziarg 2001; Collier and Hoeffler 2002; Collier 2007). The impact of ethnic diversity on economic growth processes and for that matter well-being, have been documented in many areas in the literature.

Studies have shown that high level of ethnic division tends to affect economic growth and development processes such as trust and transaction cost (Collier 2007; Leigh 2006; Knack and Keefer 1997), public good provision (Kimenyi 2006; Fosu et al. 2006; Alesina et al. 1999), contact and contracts (Bates 2000), economic growth (Easterly and Levin 1997; Montalvo and Reynal-Querol 2005; Leigh 2006) and level of investment (Mauro 1995). For example, Mauro (1995), shows that a high level of ethno-linguistic diversity has a negative effect on the level of investment. Taken as a whole, these studies suggest that social heterogeneity, especially ethnic diversity, is inimical to economic growth and development and therefore detrimental to overall well-being of people in such societies. For example, Easterly and Levine (1997) have documented that moving from an ethnically homogeneous country to one with a diversity of ethnic communities corresponded with a decrease in annual economic growth rates of more than 2 percent. They then applied this finding to Africa, reasoning that, because African countries are typically ethnically diverse, the strong link between ethnic heterogeneity and slow growth was quite likely an important part of the explanation for that region’s “growth tragedy.”

Nearly all these studies tend to employ ethnic fractionalization called ELF (Ethno-Linguistic Fractionalization) as a measure of ethnic diversity. The ELF technique of measuring ethnic diversity has been criticized because of assumption upon which it is built (see, Posner 2004; Alesina et al. 2003; Fearon 2003; Roeder 2001). Therefore, most studies on ethnic diversity and economic growth have been criticized for their measurement error and bringing conclusions from such studies into dispute (Posner 2004). Another feature of these studies is that they tend to be exclusively at the macro-level to the neglect of micro-level analysis.

Although the linkage between ethnicity and economic growth at the micro level has been alluded to in the literature, such studies tend to be descriptive rather than analytic. Some qualitative studies have argued that at the micro level, harsh economic conditions can instigate ethnic identification, heighten ethnic polarization, and inter-group competition for resources which can lead to violence, thereby, stifling economic growth. The destructive effect of ethnicity on economic growth is eloquently articulated by Schluter and Ashcroft (1988:2) as follows:

Ethnicity can be an important factor in undermining the economic growth of countries. Ethnic conflict contributes to the political instability and deterioration in law and order which in turn leads to the destruction of infrastructure, lower levels of investment and loss of skilled people.

It has been argued in the literature that when overall security and well-being of the individuals are not provided for by the state independently of their tribal affiliation, they will be encouraged to rely upon traditional tribal principles of co-operation (see Carr 1995). From this perspective, ethnic groups become the insurance people usually draw on in times of economic insecurities.

Conversely, it has been suggested that in some instances, ethnicity can serve as a catalyst for social and economic development by mobilizing people to initiate development projects in their communities. For instance, Hameso (1997:14) argues that “… in a situation of state irresponsibility, society resorts to what it knows and trust best itself.” Ethnicity assumes the role of the state through self-help. From this angle, ethnicity can serve as a catalyst for development. The positive role of ethnicity in the provision of public goods in Sub-Saharan African societies is well articulated by Osaghae (1994:11):

As government[s] ha[ve] failed to live up to [societies’] expectations, a failure which dates back to the colonial regime, the people have taken their own development initiatives in furtherance of their constitutive interest. The famous [ethnic] unions arose in the milieu, as parallel structure to provide public goods. They awarded scholarships, built schools and churches, town halls and hospitals and provided within their limits, loans to small-scale traders and artisans and engaged in other self-help projects.

The truism then is that in multi-ethnic societies, as evident in Sub-Saharan Africa, the relationship between ethnicity and economic growth at the macro-level and hence economic well-being at the micro-level is not straight forward and can be confusing at times.

3 Ethnicity and Socio-Economic Inequality in Ghana

Ghana’s population of about 22 million is characterized by ethnic diversity. The major groups are the Akan, the Mole Dagbani, the Ewe, the Ga Adangbe, the Guan, the Gurma, the Grusi and the Mande-Busanga (Ghana Statistical Service 2000). The unequal development of the country has made internal migration a common phenomenon in Ghanaian society. Due to internal migration and foreign immigration, most communities are ethnically heterogeneous, especially in urban areas. Interestingly, despite the geographic movement of people, ethnic identification and bonds continue to remain a strong feature of Ghanaian society. It is not uncommon to find ethnic enclaves in Ghanaian cities and even rural areas but sense of ethnicity and identity are paramount in rural settings which tend to be ethnically more homogenous.

Even though no region of Ghana is ethnically homogeneous, an overriding feature of the county’s ethnic polarization is the north–south divide. The southern half of the country is dominated by the Akan group. The regional and North–South disparities in opportunities have shaped and sustained variations in the well-being between various ethno-linguistic groups in the country for some time now. For example, at independence, having been largely neglected and left relatively underdeveloped under colonial rule, the northern ethnic and linguistic groups declared a social and economic distance from the rest of the country, and its political leaders argued that their people were not ready to be governed as a part of the independent Ghana without special protections. In another instance, having been colonized by Germany, and governed later as part of the UN Trusteeship Territories together with Togoland, the Ewe exerted irredentist pressures towards their cousins in Togoland. Along the same line, the Ashanti (the largest of the Akan groups) also demanded special protections for their cocoa and mineral wealth as well as their culture (Austin 1964; Boahen Adu 2000).

The difference in well-being among the ethnic-regional groups is further evidenced by the distribution of social amenities such as school, hospitals and the like. Using distribution of facilities in the country along a continuum of the most to the least developed, Dickson (1975), for example, puts forward the following list in a descending order of rank: (1) Greater Accra, (2) Ashanti, (3) Eastern, (4) Central, (5) Western, (6) Volta, (7) Brong/Ahafo, (8) Northern, and (9) Upper East and West. Using a quantitative measure ranging from 1 to 0, and using Greater Accra as the base, Ewusi (1976) ranked the regions as follows: (1) Greater Accra 1.000, (2) Cenrtal Region 0.398, (3) Western Region 0.392, (4) Eastern Region 0.355, (5) Ashanti Region 0.340, (6) Volta Region 0.306, (7) Brong/Ahafo Region 0.365, (8) Northern Region 0.110, and (9) Upper Region (East and West) 0.011.

In spite of the differences between the rank positions of the Eastern and Western, Central and Ashanti Regions, which vary in the two classifications presented by Dickson and Ewusi, respectively, there is a great deal of consistency, given that these regions are predominantly ethnically and linguistically Akan. Apart from the Brong Ahafo Region which was ranked seventh after the Volta Region in both classifications, much of the Akan areas are more developed than the Northern and Upper Regions which are ethnically non-Ewe and non-Akan. For example, data presented by Ewusi clearly shows that the index of development of the Northern and Upper Regions was 0.110 and 0.071, respectively. These are far below the indices of the others, which range from 0.265 and above. In addition, of the five regions classified as poor in Ghana in 1999, poverty levels are highest in the three Northern Regions, (the Upper East, Upper West and Northern Regions). To illustrate: nine out of ten people in the Upper East, eight out of ten in Upper West, seven out of ten in the Northern Region are poor, as compared to five out of ten in Central and Eastern Regions (Government of Ghana 2003).

Ghana has embraced the tenets of the free market with its associated liberalization policies for some time now. At the macro-level, Ghana has been able to register significant economic gains for the last two decades (Booth et al. 2004). Economic growth has averaged 4.5 percent from 1983 through 2000, but accelerated to 5.8 percent in 2004 and 6 percent in 2005 in response to the government’s program of reforms, such as structural adjustment, redeployment, devaluation of the nation’s currency, and others. Ghana has brought down poverty levels from 52 percent in 1992 to 35 percent in 2003. It is likely to surpass the Millennium Development Goal of halving poverty by 2015. Ghanaians’ access to electricity is the highest in Sub-Saharan Africa outside South Africa. Experts have suggested that Ghana is at crossroads as it is poised to achieve the middle-income country status by 2015 (Bogetic et al. 2007; World Bank 2007; Government of Ghana 2003).

The question facing Ghanaians is whether the dividends from the economic growth at the macro-level are trickling down to the people at the micro-level. Studies show that despite all the positive economic indicators at the macro-level in Ghana, people’s outlook about their overall well-being is nothing to be proud of. This is collaborated by Afro-barometer survey (2003) which reports that 85 percent of Ghanaians in the survey score their personal wealth between 0 and 5 on a scale from 0 to 10. In the same survey, nearly two-thirds (64 percent) describe their living conditions as bad, and the same number report no improvement in their standard of living during the previous 12 months. More than half (54 percent) of all respondents say they live from hand to mouth, and only 18 percent are able to save money regularly. Many Ghanaians cannot afford basic necessities of life such as food, water and medical care. Forty percent of respondents say they have gone without food, and 43 percent have gone without water, at some time during the past year. More than half (54 percent) of Ghanaians report having gone without medical attention at some time during the same period, and 39 percent did so regularly.

The general assumption that positive macro-level economic growth indicators will automatically and positively enhance people’s well-being at the micro-level has been found to be wanting around the globe for some time now. The literature is replete with evidence that faster economic growth is not always followed by faster rates of poverty reduction (Sirinivasan 2001) and hence improvement in well-being. However, some studies show that economic growth and a decline in poverty and hence better well-being run parallel (Winters 2000). There is also evidence that despite high economic growth, poverty is not reduced markedly to enhance well-being in certain cases. It is therefore argued that economic growth can enhance well-being for the poor only if additional measures are instituted towards the poor. This is often termed pro-poor growth measures (Streeten 1971), and this is what is required in Ghana today.

The only way to promulgate an effective pro-poor program is to understand the factors shaping the well-being of people at the micro-level. At micro-level, one factor which tends to be consigned to the periphery in well-being discourse is the importance of ethnicity in molding people’s well-being. In multi-ethnic society like Ghana, the salience of ethnicity to many people, as well as the extent of their attachment to their ethnic values, practices and expectations, enables ethnicity to be influential within a wide range of domain of life experiences and decision making. Previous studies using data from Ghana have established that ethnicity informs behaviors within various realms including political behavior (Agyeman 1995), educational attainment (Sackey 2005), reproductive health decisions (Addai 1999a), and sexual behavior (Addai 1999b), among others. The general belief by the authors is that the relevance of ethnicity in the day-to-day live of Ghanaians touches on every aspect of life and that economic well-being cannot be an exception. Against this background, the logical question becomes “what is the relative importance of ethnicity in people’s day-to-day economic well-being?

Although ethnicity has been implicated in various studies by economists and political scientists at the macro-level in Sub-Saharan Africa, little is known about the extent to which ethnicity influences economic well-being of people at the micro-level, and more importantly in societies such as Ghana which has shown economic growth for decades. This is an unfortunate omission in this area of research as success of any economic program is gauged by the extent to which it transforms people’s well-being at the micro-level. The basic question raised in this paper is “what is the relative effect of ethnicity in shaping economic well-being (perceived and actual) among Ghanaians at the micro-level after decades of economic growth? Put directly “does ethnicity instigate or undermine economic well-being (perceived and actual) at the micro-level in the context of positive economic growth in Ghana? Understanding ethnic differences in perceptions of economic well-being may be as important as examining ethnic differences in actual economic well-being in the politically and ethnically charged environment which characterizes a society like Ghana.

Unlike previous studies that tend to use ELF index of ethnicity, this study draws on respondents’ own reported ethnic identification as a measure of ethnicity to probe the relationship between ethnicity and economic well-being (perceived and actual) among Ghanaians. If ethnicity per se emerges as a significant predictor of people’s economic well-being (perceived and actual) for a particular group(s), then the norms, practices, beliefs, expectations and other elements of such a group(s) can be tapped for betterment of people’s well-being. Although we acknowledge there are no direct measures that allow us to assess how specific beliefs, norms, values, and practices may instigate or undermine individual’s economic well-being, we use the proxy indicator of ethnic identification—sethnicity to arrive at this goal.

4 Data and Methods

We examined the links between ethnicity and economic well-being in Ghana by analyzing data from the 2008 Afro-barometer study conducted in Ghana (n = 1200). This survey was sponsored by Michigan State University, the Institute for Democracy in South Africa, and the Centre for Democratic Development in Ghana and is one of the series of surveys that have been conducted in a sample of African countries since 1999. As in the other participating countries, the Ghana sample is nationally representative of the adult population (i.e., those over 18 years old and eligible to vote). Among other things, the survey collected detailed information about the respondents’ views about democracy, their participation in the electoral process, governance, livelihoods, economic concerns, social capital, conflict and crime, and perceptions about national identities.

In addition to the above, detailed information on the respondents socio-demographic characteristics, including their age, education, place of residence, and religious orientations and involvement and ethnic identification were also collected. It is the latter questions plus those dealing with economic well-being that we use in this study. Economic well-being is measured by self reported assessment of living conditions of various ethnic groups in the country (perceived well-being) as well an index of actual economic well-being. The distribution of reported economic well-being and actual economic well-being among the ethnic groups will be assessed. The kind of relationship between ethnicity and perceived and actual economic well-being will be examined. Given the differential impact of social and economic development in the country, analysis will also assess how ethnicity affects people’s economic well-being in terms of the urban–rural setting and age cohort (15–34; 35 and up).

5 Analytic Techniques and Dependent Variables (DV)

The analysis is carried out at three stages. The first stage uses a non-parametric method (Spearman’s rho) to look at the relationship between each variable with our dependent variable(s). Spearman’s rho (p) is also known as Spearman’s rank correlation coefficient and is informative of the direction and strength of a bivariate relationship. This method was employed generally because of the predominantly categorical (ordinal) variables in our study. Spearman’s rho and Kendall’s tau are prominent methods used for correlating these types of variables (Bryman and Duncan 2009).

In the second stage, logistic regression equations are used to estimate multivariate models of the odds or the likelihood of a respondent reporting better perceived economic well-being rather than worse, while taking into consideration their ethnic backgrounds. The model helps to estimate perceived economic well-being for different ethnic groups while simultaneously controlling for other measurable factors associated with economic well-being. Perceived economic well-being is defined in terms of the respondent’s perception of his/her ethnic group’s economic well-being compared to other ethnic groups. We recoded the responses to the original question (1 = much better to 5 much worse) into a binary measure, with 1 = better (including “much better” and “better”) and 0 = worse (combining “same”, “worse” and “much worse”). If ethnicity has an impact on perceived economic well-being, then when other background variables are controlled for by logistic regression, the coefficients for ethnicity should be significant. The logistic regression model estimates a linear model in the following form:

$$ \ln ({\text{pi}}/[1 - {\text{pi}}]) = {\text{bo}} + {\text{biXi}} $$

where pi is the estimated probability of a particular event occurring to an individual with a given set of characteristics. Xi; bo is a constant that defines the probability Pi for an individual with all Xi set to zero; bi are the estimated coefficients. The ratio pi/[1 − pi] is the odds ratio of respondents with given set characteristics reporting better versus worse perceived economic well-being. The estimate of bi for a particular covariate Xi is interpreted as the difference in the predicted log odds between those who fall within that category of characteristics and those who fall within the reference or omitted category for that characteristic. If each estimated bi is exponentiated (Exp[bi]), the result can be interpreted as giving the relative odds of having better perceived economic well-being for those individuals with characteristic Xi, relative to those individuals in the reference group. All results of multivariate models are given as the exponentiated coefficients. Overall, we estimated ten models to assess the relative contributions of ethnicity and non-ethnic factors in the likelihood of having better or worse perceived economic well-being.

The first model is applied to the whole sample and uses only the independent variable (ethnicity) as a predictor of the probability of falling into the “better” or “worse” category. The second model incorporates all the controls in addition to the independent variable. This same sequence is used for the remaining eight models, in that Models III and IV are capturing only the urban sub-sample and Models V and VI only the rural sub-sample. Models VII and VIII capture only the older age group and Models IX and X focus on the young age group. Zero category of predictor variables with more than two categories is used as a reference category, that is, the simple-first method of contrast is applied. Four significance levels are considered: .10, .05, .01, and .001.

The third stage incorporates multiple regression equation with categorical variables technique. Being that our second dependent variable (actual well-being) is quantitative and our independent variable and covariates are categorical (with two or more categories), we derived a series of dummy variables (16) out of those categorical variables with more than two categories (ethnicity, region, religion, and education) to make them suitable for multiple regression. The reference categories for these variables correspond to those from logistic regression.

Ten models, in the same sequence as for logistic regression, are formed and the same significance levels are considered. The dependent variable (actual economic well-being) is composed of 5 items (see Appendix) which were summed to produce the actual economic well-being index varying from 0 to 5 (Cronbach’s Alpha = .806). The 5 items originally indicated the frequency of going without food, water, medical care, cooking fuel, and cash income (0 = never, 4 = always) over the previous year for each respondent or anyone in his/her family. These items were chosen as they represent the very basic and indispensable sources or measure of actual (objective) economic well-being. The original variables (items) were recoded (0 = never, 1 = once or more) and a measure of actual well-being was composed by counting every 1 on any of these items.

The smaller the number, the positive or better ethnicity is on well-being. Smaller numbers mean going without fewer important items over the past year. Consequently, in multiple regression analyses, if there is a significant negative coefficient, it would mean that a certain predictor has a decreasing effect on the number of actual well-being items gone without, which is more favorable (fewer items lacked once or fewer times over the past year) than if it has an increasing effect (positive coefficient). Positive effect means lacking more of the five basic items in the actual well-being index. Therefore a positive coefficient will be interpreted as ethnicity having negative effect on well-being (going without more of the items) and negative effect will suggest ethnicity has positive effect on well-being (going without less of the items). This may be counter-intuitive at first, but the essential part is that a negative relationship means better actual economic well-being and vice versa.

6 Independent Variable

The independent variable for the study is ethnic affiliation categorized into the following major groups: (0) Other Ghanaians; (1) Akans; (2) Ewe/Anglo; (3) Ga-Adangbe, and (4) Dagomba.

7 Control Variables

Control variables considered in our multivariate models include:

  1. (a)

    Respondent’s age (coded as: 0 = 15–34, 1 = 35 and up)

  2. (b)

    Respondent’s sex (coded as: 1 = male, 0 = female)

  3. (c)

    Respondent’s educational attainment (coded as: 0 = No Education, 1 = Elementary, 2 = High School, 3 = Bachelor Degree and Higher).

  4. (d)

    Respondent’s employment status (coded as: 1 = Employed, 0 = Not Employed).

  5. (e)

    Respondent’s place of residence (coded as: 1 = Urban, 0 = Rural).

  6. (f)

    Respondent’s region (coded as: 0 = Western/Central, 1 = Greater Accra, Volta, 3 = Eastern/Ashanti/Brong Ahafo, and 4 = Northern/Upper East and West).

  7. (g)

    Respondent’s religion (coded as: 0 = No/Traditional, 1 = Other Christians. 2 = Catholic, 3 = Protestant, 4-Muslim, 5 = Evangelical/Pentecostal).

Each predictor of economic well-being has been selected for inclusion in the analysis for explicit theoretical reasons and availability in the data set. For a host of reasons like access to employment and reliable sources of income, level of educational attainment is expected to be conducive to economic well-being.

8 Results

8.1 Bivariate Correlations

Table 1 depicts percentage distributions of all variables used in the analyses (on the first DV = perceived well-being), and Spearman’s rho coefficients and their significances are included for each bivariate relationship. For the first set of analyses pertaining to perceived economic well-being, ethnicity was significantly correlated with it (ρ = .060, p < .05). The remaining variables that were significantly correlated with perceived well-being are: education (ρ = .108, p < .001); residence (ρ = .105, p < .001); and region (ρ = .134, p < .001).

Table 1 Percentage distribution of all characteristics between those reporting better or worse perceived well-being: Ghana, 2008 (N = 1200)
Full size table

Table 2 shows mean distributions of each predictor variable category on the second DV (actual well-being) and bivariate correlation coefficients (ρ). For the second set of analyses concerning actual economic well-being, the following variables were significantly and negatively correlated to it: education (ρ = −.190, p < .001); employment (ρ = −.056, p < .10); residence (ρ = −.127, p < .001); region (ρ = −.057, p < .05); and religion (ρ = −.064, p < .05). All the correlations reported here are statistically significant. For example, education is significantly and negatively correlated with actual well-being. The mean of actual well-being values for those with no education is the highest (2.44) relative to other education category. This illustrates that those with no education lacked more items once or more times in the past year than those with a BA or higher degree who lacked the least (1.51). Actual economic well-being is not significantly correlated with ethnicity.

Table 2 Mean distribution of all characteristics on actual well-being (going once or more times without 0–5 items over past year): Ghana (N = 1200)
Full size table

8.2 Logistic Regression Results

Table 3 presents logistic regression analyses results. In Model I (entire sample with ethnicity predictor only), the results reveal that three ethnic groups are more likely to report better perceived economic well-being than Other Ghanaians. Whereas Akans and Dagombas are significantly approximately three times more likely to report better perceived economic well-being compared to Other Ghanaians, the Ga-Adangbe are about 2 times more likely to do the same. In Model II, ethnicity and all the remaining variables are introduced. Two ethnic groups compared to Other Ghanaians remain significantly more likely to report better perceived well-being: Akan (Exp[B] = 2.67, p < .001) and Dagomba (Exp[B] = 3.08, p < .001).

Table 3 : Odds ratios [Exp(B)] of reporting better versus worse perceived well-being: Ghana, 2008
Full size table

Focusing on control variables, the odds of reporting better perceived well-being tend to be 46 percent lower for those with elementary education in comparison to those with no education (Exp[B] = .54, p < .05). The odds of reporting better perceived well-being tend to be higher among residents of three regions: Greater Accra (Exp[B] = 1.81, p < .05); Eastern/Ashanti/Brong Ahafo (Exp[B] = 3.35, p < .001); and Northern/Upper East and West (Exp[B] = 3.18, p < .001) than those living in the Western/Central region. The data reveal that the probability of reporting better well-being tends be significantly higher for Other Christian (Exp[B] = 2.12, p < .05); Catholic (Exp[B] = 1.78, p < .10), and Protestant (Exp[B] = 2.24, p < .01) groups than the reference group (No and Traditional Religion group).

Model III (urban sub-sample, ethnicity predictor only) shows that three ethnic groups significantly influence how people perceive their groups well-being in urban areas. It is interesting to note that, while the probability of reporting better outlook on well-being tends to be almost 2.5 times higher for two ethnic groups (Akan (Exp[B] = 2.43, p < .001) and Dagomba (Exp[B] = 2.46, p < .05) than Other Ghanaians in urban setting, the Ewe/Anglo urban residents are 43 percent significantly less likely to perceive their well-being in the same way (Exp[B] = .57, p < .10). Even with the introduction of controls (Model IV), the positive relationship between Akans and Dagombas ethnic background and perceived well-being continues to persist in the urban setting. These groups are three times more likely to project better well-being compared to their Other Ghanaian counterparts in urban areas. Whereas the significance of Ewe/Anglo background on well-being vanishes upon controls, interestingly, another ethnic group (Ga-Adangbe) emerges as a significant predictor of perceived well-being in urban settings. The likelihood of perceiving better well-being is 2.5 times higher among Ga-Adangbes residing in urban areas compared to their counterparts in the Other Ghanaian group living in the same settings.

In Model IV, the data show that non-ethnic factors play a significant role in molding Ghanaians’ perception of their group’s well-being in urban areas. For instance, the odds of having a better outlook on one’s ethnic well-being tends to be 75 percent lower for urban residents with elementary education in comparison to those with no education (Exp[B] = .35, p < .05) in urban settings. However, for an urban dweller in two regions (Eastern/Ashanti/Brong Ahafo (Exp[B] = 4.22, p < .001) and Northern/Upper East & West (Exp[B] = 4.25, p < .01), the probability of reporting better perceived well-being tends to be more than 4 times higher than that of their counterparts in Western/Central. It is equally important to note that the odds of perceiving better well-being among urban residents in Greater Accra tend to be 2 times higher than the odds for their counterparts in the reference group.

Models V and VI reflect the rural sub-sample only. In Model V, the results show that rural residents belonging to Akans, Dagombas, and Ga-Adangbes ethnic groups demonstrate a higher tendency to perceive better well-being than their counterparts in the Other Ghanaians ethnic group. For instance, the likelihood of members of Akans and Dagombas ethnic groups residing in rural areas to report better well-being tends to be approximately 3 times higher than that of Other Ghanaians. In Model VI, just as in previous models, the two ethnic groups (Akans and Dagomba) remain to be more likely to perceive better well-being than the reference group. Also, in the rural setting, those living in Eastern/Ashanti/Brong Ahafo and Northern/Upper East & West regions tend to be about 3 times more likely to report better perceived well-being than those living in the rural areas of the Western/Central region. The odds of Other Christians and Protestants in rural areas reporting better well-being tend to be about 2.5 times higher than those with No and Traditional religion in rural areas.

Models VII and VIII capture the odds of reporting better versus worse perceived economic well-being among the 35+ age cohort. In Model VII, older Akans (Exp[B] = 2.48, p < .001) and older Dagombas (Exp[B] = 2.19, p < .05) are twice more likely to report better well-being than older Other Ghanaians. Belonging to the Ewe/Anglo ethnic groups tends to significantly induce negative perception of well-being among the older cohort. The odds for reporting worse perceived economic well-being tends to be 54 percent significantly higher (Exp[B] = .46, p < .05) among older Ewe/Anglo people compared to older Other Ghanaians. Model VIII shows that even after the necessary controls, Akan and Dagomba backgrounds continue to instigate significant likelihood of reporting better perceived economic well-being among the old age cohort.

In the same older cohort, structural factors play a vital role in shaping people’s view of well-being. For example, the older cohort members with elementary education are 67 percent less likely to report better perceived well-being (Exp[B] = .43, p < .05) than their counterparts with no education. The data suggest that for older Ghanaians living in Greater Accra, Eastern/Ashanti/Brong Ahafo, and Northern/Upper East & West regions the probability of significantly perceiving better well-being tends to be higher than for the reference group. Finally, being affiliated with four religious faiths (Other Christian, Catholic, Protestant, and Evangelical/Pentecostal) tends to induce significantly higher odds (2.42–3.39) of reporting better perceived well-being compared to those who are affiliated with No and Traditional Religion among older Ghanaians.

The remaining two models reflect the younger age cohort sub-sample. In model IX, young Dagombas, Akans, and Ga-Adangbes are about 4 times, 3.5 times, and 2.5 times, respectively more likely to report better well-being in comparison to young Other Ghanaians. Model X incorporates all the controls and uniquely shows that all ethnic groups are significantly more likely to report better perceived well-being than Other Ghanaians among the young cohort. Young Ghanaians living in the Eastern/Ashanti/Brong Ahafo and Northern/Upper East and West regions are also significantly more likely to report better perceived well-being than their counterparts in the reference group.

8.3 Multiple Regression Results (OLS)

Table 4 presents multiple regression analyses results. The sequencing of samples/sub-samples and variables in the models is the same in multiple regression with dummies as in logistic regression. Model I shows that being Akan and Ga-Adangbe significantly reduces the number of lacked actual well-being items once or more times in the previous year than being a member of Other Ghanaian ethnic group (β = −.189, p < .001 and β = −.085, p < .01, respectively). In other words, belonging to Akan and Ga-Adangbe is significantly related to better actual well-being compared to membership in the reference group. In Model II, net of controls, the relationship between well-being and Akan and Ga-Adangbe background vanishes and, interestingly, a new ethnic group (Ewe/Anglo) emerges as significantly shaping actual well-being. The analyses reveal that Ewe/Anglo ethnic background negatively impacts actual well-being of Ghanaians.

Table 4 : Total actual well-being items gone without once or more times past year regressed on selected characteristics: Ghana, 2008
Full size table

Focusing on non-ethnic factors, all levels of education significantly induce better actual economic well-being compared to those with no education. Similarly, living in urban areas significantly instigates better actual well-being than living in rural areas (β = −.051, p < .10). Focusing on region of residence shows that only living in the Eastern/Ashanti/Brong Ahafo region significantly induces better actual well-being compared to the reference group (β = −.265, p < .001). The data suggest that being a member of Protestant, Muslim, and Evangelical/Pentecostal faith significantly instigates better actual well-being than belonging to No and Traditional Religion.

In Model III, only Akan ethnic background induces a significantly better effect (β = −.150, p < .05) on actual well-being in urban settings. It is interesting to note that even after controls (Model IV), the positive relationship between Akan ethnic background and actual well-being continues to persist in urban areas. Unlike the previous model, attaining elementary education is not significantly related to actual economic well-being in urban areas. For a person to enjoy better well-being in urban areas from his or her education compared to those with no education, one must have at least a high school diploma (β = −.227, p < .01) and a university degree (β = −.227, p < .01). In Ghana’s urban setting, one’s actual economic well-being is significantly molded by region of residence. The data suggest that living in Greater Accra, Northern/Upper East & West and especially Eastern/Ashanti/Brong Ahafo significantly instigates better actual well-being than living in Western/Central region, among urban dwellers. One’s faith is also strongly related to actual economic well being in urban areas. The data reveal that membership in Catholic, Protestant, and Muslim denominations significantly induces better actual well-being.

Models V and VI reflect the rural sub-sample. Model V shows that being Akan and Ga-Adangbe significantly decreases the number of deprivations in actual well-being that one has experienced once or more times over the past year (β = −.170, p < .001; and β = −.118, p < .01, respectively) in rural areas. This means that in rural settings, belonging to the two ethnic groups (Akan and Ga-Adangbe) predicts better actual economic well-being compared to their counterparts belonging to the other Ghanaian group.

In Model VI, upon controls, the number of ethnic groups significantly predicting actual economic well-being increases from two groups to three. Ewes/Anglos join the Akans and Ga-Adangbes as groups whose background significantly shapes actual economic well-being in rural areas. The interesting observation here is that the relationship between ethnicity and actual well-being changes from positive to negative among Akans and Ga-Adangbes. Thus, in rural settings, controlling for non-ethnic variables in the equation, membership in Akan, Ga-Adangbe, and Ewe/Anglo ethnic groups induces a higher chance of lacking more basic necessities of life. Thus, the data suggest that, in rural Ghana, membership in Akan, Ga-Adangbe, and Ewe/Anglo ethnic groups tends to undermine the actual economic well-being of members of these groups compared to the reference group.

Just like at the national level, all levels of education significantly instigate better actual economic well-being than no education in rural settings. Also, living in Greater Accra and Eastern/Ashanti/Brong Ahafo regions significantly predicts better actual economic well-being than living in the Western/Central region in rural areas. The same cannot be said for those living in Northern/Upper East & West regions. Among rural residents living in Northern/Upper East and West significantly predicts worse actual economic well-being.

Among the older age cohort (Model VII), the results show that membership in Ewe/Anglo ethnic group undermines actual economic well-being (β = .120, p < .01). Model VIII shows a somewhat similar result like Model VI, that is, three ethnic groups are significantly and positively related to actual economic well-being. Hence, these groups (Akan, Ewe/Anglo, and Ga-Adangbe) predict worse actual well-being than older Other Ghanaians, net of all the controls just as in Model VII. Analogous to the urban setting result, high school and BA or higher education level significantly predict better actual well-being compared to the no education category, but the same cannot be said for elementary education. For older Ghanaians, living in Eastern/Ashanti/Brong Ahafo region positively impacts actual economic well-being (β = −.302, p < .001). Finally, being a member of Protestants and Evangelicals/Pentecostals religious faiths minimizes the risk of not having some basic necessities and therefore induces significant improvements in actual economic well-being among the older Ghanaians.

Models IX and X report the findings among the younger sub-sample. Model IX demonstrates that the Akan and Ga-Adangbe ethnicity (β = −.312, p < .001; β = −.134, p < .01, respectively) significantly leads to better actual economic well-being. Model X includes control variables and reveals that Akan background continues to have a significant effect on better actual well-being (β = −.135, p < .10). Having elementary and high school education (but no BA or higher) significantly induces better actual well-being than having no education. Living in Eastern/Ashanti/Brong Ahafo also significantly predicts better actual economic well-being than Western/Central (β = −.211, p < .001).

9 Discussion

The bivariate analysis reveals that there is a significant relationship between ethnic background and how people perceptive the well-being of ethnic groups. However, there is no significant relationship between ethnicity and actual well-being. Thus, people’s perception of ethnic economic well-being does not correspond with how people in various ethnic groups are actually doing economically at the micro-level. The reported ethnic differences in perception of well-being may be attributable to people’s sense of historical inequality which has existed in the country from the colonial times until today.

Our multivariate results indicate that ethnicity is an important determinant in how people perceived the well-being of their ethnic group. For Akans and Dagombas, ethnicity significantly instigates better perception of well-being in national, urban, and rural settings as well as among young (15–35) and old (35 and up) age cohorts. This is an interesting finding in that the two ethnic groups have nothing in common in terms of economic opportunities. Whereas the Akans mostly reside in the relatively developed part of the country, South, the Dagomba’s dominate the least developed part of the country, North. The Akans’ positive outlook on well-being may be attributable to the relative advantage the group enjoys in the country due to access to social amenities and employment opportunities. Dagombas’ positive outlook on well-being may be due to recent development efforts by the government and non-governmental organizations to make a difference in the Northern part of the country where the Dagombas dominate. For example, the Northern part of the country now has a university and investors are been encouraged to invest in this part of the country.

Although ethnicity significantly induces better outlook of well-being among Ga-Adangbe and Ewe/Anglo ethnic groups, these positive effects of ethnicity prevail only among the young age cohort. This finding may be due to exposure to modern ideas and technology among the young age cohort. Interestingly, whereas ethnicity induces positive perception of well-being among rural Ga-Adangbes, it promotes negative perception of well-being in urban settings among Ewe/Anglos.

Consistent with the existing literature, region of residence emerges as the most important non-ethnic predictor of perceived well-being in Ghana. In all settings and among all groups, living in Eastern/Ashanti/Brong Ahafo tends to lead to better perception of economic well-being. The impact of region on the perception of well-being prevails at the national level, among urban dwellers, as well as among the older cohort for Greater Accra residents. These findings may reflect the relative advantage these regions enjoy in terms of access to opportunities and socioeconomic amenities in the country. Contrary to the expectation, living in the Northern/Upper East, & West predicts positive outlook on economic well-being at all levels and among all groups. This is an interesting finding as these regions are the least developed in the country and deserves to be further studied,

As expected, education is an important determinant of well-being in Ghana. Education is an important non-ethnic factor that tends to mold people’s outlook on economic well-being. Interestingly, it is shown that attaining just elementary education tends to induce less hope among Ghanaians about their well-being at the national level, in urban areas, and also among the older cohort. This is a peculiar finding in that people’s perception of their economic well-being has serious implications for the stability of the nation. It is greatly on education to induce some sense of hope so as to forestall any social instability in the country. This finding may reflect the reality in the country that higher education above elementary offers a chance of economic mobility in the country.

Religion is a significant determinant of how people of Protestant and Other Christian faiths perceive well-being in Ghana. The relationship between religion and better perceived economic well-being as reported among Other Christians and Protestants may be explained in terms of religious teachings these religious faiths espouse. Instead of the traditional salvation message preached by Orthodox churches, the new sectarian churches (Other Christians) and the Pentecostals now preach what has come to be known as “prosperity” message. These churches basically entreat their members to shun away from negativity about life and always project positive outlook about life. This teaching may be shaping their mindset and hence their perception about well-being in the country. An appealing observation is that among the older cohort, with the exception of Muslim members, religious background tends to induce significant better outlook about economic well-being. Even here, the odds of reporting better perceived well-being tends to be higher among the older Pentecostals and Other Christians.

Looking at actual economic well-being, the findings from this study suggest that ethnicity is a salient determinant of people’s well-being in Ghana. Among Akans, ethnicity is both a curse and a blessing as far as actual well-being is concerned. Whereas ethnicity minimizes the risk of going without basic essentials among young people and urban dwellers, it tends to increase the risk of going without basics among rural residents and the older cohort. For members of Ewe/Anglo ethnic group, ethnicity seems to undermine well-being at the national level and in the rural setting, as well as among older people. Finally, among Ga-Adangbes, ethnicity undermines well-being among rural dwellers and also among the older cohort.

With the exception of Dagombas, ethnicity significantly undermines the economic well-being of all remaining ethnic groups under study (Akans, Ewe/Anglos and Ga-Adangbes) in rural areas and also among older people. This is a very significant finding. Majority of Ghanaians live in rural areas so the impact of ethnicity on well-being is affecting the majority of citizens in the nation. The negative impact of ethnicity on economic well-being in rural areas among the ethnic groups and also older people may be due to traditional practices, value and expectations. The importance of traditionalism shaping well-being may be supported by the fact that rural dwellers and older group are likely to have high adherence to traditionalism. For example, the older and rural population may be less willing to embrace new ideas and also trust strangers in transactions and therefore hold onto traditional practices and values which may undermine their economic well-being.

As expected, education tends to be the most important determinant of a person’s actual economic well-being in the country. All levels of education seem to minimize the risk of going without basics in Ghana. This is due to the fact that education is the main engine of mobility in Ghana. Having education offers one the chance of getting a job and being employed in the country. Hence education serves as insurance against economic challenges in the country. Therefore having a higher level of education logically minimizes the risk of not being able to meet one’s basic needs. Living in Eastern/Ashanti/Brong Ahafo tends to show consistent reduction in the chance that someone from these regions will go without basic necessities at all settings and among all groups. Northern/Upper East and West region shows an inconsistent impact on well-being. Interestingly, while it minimizes the risk of going without basics among rural residents, the opposite is the case among urban dwellers. This may be due to a lack of jobs in urban areas. The rural social organization seems to fight against deprivation of basic necessities in the region whereas the same cannot be said of the structures in urban settings. Religion does not seem to play an important role in the people’s actual economic well-being.

10 Findings and Conclusions

While there is a good deal of knowledge about the connection between ethnicity and economic growth at the macro-level, the same cannot be said about the interplay between ethnicity and economic well-being at the micro-level. Our review of the literature reveals that little, if any, existing research has systematically examined the impact of ethnicity on economic well-being at the micro-level. To help fill this vacuum, we made it our purpose in this paper to assess the influence of ethnicity on economic well-being. We employed two measures of economic well-being—perceived and actual—in three settings (national, urban and rural) and among two age groups (15–34 and 35 and up). As it turns out, ethnicity matters for different measures of well-being and within different settings and among different groups.

This study shows that after decades of economic growth in Ghana, ethnicity continues to be a significant determinant of people’s perception of and actual well-being at the micro-level. For two ethnic groups (Akans and Dagombas) the analysis reveals that ethnicity has a significant positive effect on perception, or what one might also call psychological effect on well-being. When we consider actual economic well-being, ethnicity emerges as a critical variable in the quality of life in Ghana. However, the impact of ethnicity on actual well-being is not straight forward as it instigates or undermines one’s well-being among Ghanaians in various settings and among various groups. The study suggests that after decades of successful economic reforms in Ghana, ethnicity has not lost its hold on how people perceive their ethnic group’s well-being and its impact on quality of life in Ghana. This is particularly the case among the majority of rural dwellers. Taken as a whole, there is a significant relative effect of ethnicity on economic well-being (actual and perceived) at the micro-level however the direction of that impact varies from population to population (sub) groups.

Findings from this study suggest that to continue to pursue the economic reforms which started decades ago without compromising the stability of the country, demands addressing how people perceive their ethnic group’s economic well-being. It is important for people to come to terms with the realities of their group’s economic well-being. The data suggest that there are no significant correlation differences between various ethnic groups when we focus on actual well-being, but people’s perception of group well-being is a different story. In a region known for ethnically induced violence and instability because of perceived economic inequalities and marginalization, the negative relationship between ethnicity and perceived well-being among some ethnic groups needs to be addressed with all the seriousness and resources available.

The salience of ethnicity in molding well-being in Ghana suggests that this variable needs a critical consideration in policy making in the country. This calls for group-specific studies to understand aspects of the various ethnic groups that tend to increase or decrease well-being. Achieving better well-being or quality of life for all Ghanaians requires that we understand how norms, beliefs, values, expectations and practices of each group shape members’ perception and actions at the micro-level. For programs to make the maximum impact on the lives of Ghanaians, they must be ethnic-specific in outlook. Instead of the “one size fits all” developmental approach in the past, which the country has little to show for at the micro-level, ethnic-specific programs stand a better chance of making a difference in people’s lives. This is because such programs will reflect people’s values, norms, beliefs and world outlooks about economic activities. As demonstrated in this paper, ethnicity is an essential part of people’s mindset, as well as a determinant of actual well-being in the country (from multivariate models) and therefore needs to be incorporated into any development strategy that has any hope of achieving desired goals.

Policy makers and leaders of Ghana need to recognize the constraints posed by ethnicity at both macro-level and micro-level of the country and to face them realistically. Any program that promises quick results for the country without considering the ethnic factor will likely only raise hopes and fail conspicuously. Such an outcome is likely to undermine the progress that the country has already registered. The truth then is that economic reform initiatives can improve well-being for majority of Ghanaians only if they are envisioned for a realistically long-range horizon, implemented consistently and comprised of sensible, politically sustainable measures after taking into account the role of ethnicity.

The study convincingly suggests that besides ethnicity, the best investment that Ghana can make to enhance individual’s quality of life is to invest in education. Education has far-reaching implications as far as people’s outlook about well-being is concerned. This is an important investment as people’s psychological mindset can make or break a country. If people have a positive outlook and hope about their life chances, then the likelihood of engaging in social unrest and destructive behavior which can destabilize the country is very minimal. Regional inequality equally needs attention in the country’s developmental efforts. It is an obvious reality that the well-being of Ghanaians will be highly enhanced if various regions receive the same attention and resources as the capital region- Greater Accra. The extreme disparities between the regions, in terms of basic infrastructures such as roads, schools, clinics and access to drinkable water, are not good for the nation as potential benefits to the nation are lost due to poor infrastructure and wasted human resources.

In a nut shell, ethnicity is a double-edge sword as it can serve as a source of blessing, but in certain situations it can instigate chaos and undermine economic growth and people’s well-being. Therefore, ethnicity deserves a serious critical look in developmental plans, strategies, and also in efforts towards the economic betterment of people’s well-being. In the mist of the on-going global crisis, improving the well-being of Ghanaians will require a multi-dimensional approach focusing on both external and, especially, internal factors of which ethnicity needs to take a center stage. Ethnic-specific approach to development in Ghana can easily be built into the decentralization processes going on in the country. Ethnic-specific programs have a chance of enhancing the well-being of the majority of the people if planned at the local level, taking into account values, norms, beliefs, practices, and expectations of the communities. The negative effect of ethnicity on well-being among the rural majority suggests that unless various districts in country adopt ethnic-specific development strategies, the chances of improving the well-being of the majority of Ghanaians will continue to elude politicians and policy makers.