The determinants of institutional trust in Botswana ’ s liberal democracy

Botswana is considered as a hub of good governance, and one of the least corrupt countries in Africa. Yet empirical evidence based on the Afrobarometer perception surveys from 2008 to 2014 suggests a decline in institutional trust. This study uses the 2014 Afrobarometer survey to explain trust in four political institutions namely the presidency, the ruling party, parliament and local council authorities. Theories of institutional trust suggest that trust is linked to performance of institutions on a number of key factors. But for the purposes of this study, we explain trust by perceptions on corruption, democracy, civic participation, government performance, level of education, age and location. The study finds that the level of education, perceptions on government performance, corruption and satisfaction with democracy are important in explaining trust in political institutions. However, safe for communing together to raise issues, civic participation is not important in explaining institutional trust. The argument of the study is that even though Batswana do not have a culture of civic engagement, they are critical in government performance, democracy and corruption.


INTRODUCTION
Botswana is regarded among the best performing African countries in good governance, and rated as Africa"s least corrupt country.The country has recorded the fastest economic growth in the world in the 1980s, and has been a frontrunner in democratic practice on the African continent evidenced by conduct of free and relatively fair elections.
Even though there has yet to be an alternation of power between parties, respect for term limits and smooth transition of power between leaders, albeit of the same party, have been consistent in a continent characterised by unconstitutional transitions and extension of presidential term limits.Botswana continues to attract positive ratings from the Mo Ibrahim Index of Good Governance and has consistently been ranked as Africa"s least corrupt country by Transparency International.
But on the other hand, empirical evidence from the Afrobarometer survey depicts a decline in institutional trust from 2008 to 2014.Therefore, the purpose of this study is to examine factors that explain trust in political institutions.The study is based on round six of the Afrobarometer perception survey that was conducted in 2014.Although a decline in perceptions of institutional trust is observed from 2008 to 2014, the present study is concerned with modelling the possible determinants of trust in political institutions in 2014.Models that explain the declining trends in institutional trust are subjects for the next research.
This study observes that the literature on determinants of institutional trust is acutely lacking in the context of Botswana.So to the best of this study knowledge, the study represents the first attempt to model determinants of institutional trust in Botswana with selected variables.For the purpose of this study, political institutions refer to those institutions that citizens select officials to represent them, namely the presidency, parliament, local government council and the ruling party.The study focuses on these four important institutions because as it is later shown that public confidence is important for the consolidation of democracy.
The study found out that institutional trust is explained by perceptions on government performance in delivery of services, political representatives involved in corruption and satisfaction with democracy.Educated people are less likely to trust political institutions than people without formal education while civic participation does not explain institutional trust.Based on these findings, the central argument of this study is that while Batswana do not have a culture of civic engagement, they are increasingly becoming critical of their government and democracy because of perceived poor government performance, perceived corruption in institutions and their dissatisfaction with democracy.

Institutional trust
It has now become generally accepted that institutional trust is an important ingredient for any functioning democracy.Mishler and Rose (2001) plainly state that trust is critical to democracy, and Bianco (1994) similarly points out that trust links ordinary citizens to the institutions that are intended to represent them, thereby enhancing both the legitimacy and the effectiveness of democratic government.When people trust their institutions, they have confidence that those institutions perform in accordance with their expectations or at least account in the event of non-performance.This is especially essential for democracies as it is a part of the social contract between elected political representatives and voters.Catterberg and Moreno (2005) argue that trust is especially important for democratic governments since they cannot rely on coercion to the same extent as other regimes.This implies that there is more support for democracy where citizens have faith in political institutions.A number of studies have also argued that a public"s trust in the actors and institutions of political authority facilitates democratic consolidation in that institutionally-trusting individuals have been found to be more supportive of democratic principles (Seligson and Carrión, 2002).Newton (2001) similarly points out that political trust is essential for democratic and stable political Seabo and Molefe 37 life.
Institutional and cultural theorists trace the origins of trust and offer varying perspectives on this issue.According to cultural theories, trust is exogenous, that it originates outside the sphere of politics in long-standing beliefs about people emanating in cultural norms, learned through process of socialization (Mishler and Rose, 2001).Cultural theorists emphasize that institutional trust is an extension of interpersonal trust which is learned early in life, and projected into political institutions.So political trust is based on attitudes and values that are learned early in life, and are transmitted from generation to generation (Inglehart, 1997;Putnam, 2000).
On the other hand, institutional theorists argue that political trust is a result of expectations by people from institutions to perform well.In this vein, Mishler and Rose (2001) posit that "trust in institutions is rationally based; it hinges on citizen evaluations of institutional performance."For Wuthnow (2002), "institutional trust which is confidence in institutions, points to the fact that much of contemporary life depends less on informal, interpersonal transactions than on the norms and social structures in which these specific transactions are embedded."From a glance, the two theoretical traditions are mutually reinforcing because trust earned through socialization is indeed later translated into institutional trust.In fact, according to Blind (2006) institutional scholars have come to accept that culture conditions institutional performance.But institutional performance is based on a number of factors which affect and determine people"s confidence on institutions.Before examining these determinants of institutional trust, a brief definitional exercise of the concept of institutional trust is in order.For purposes of this paper, institutional trust is used interchangeably with political trust because the paper is based on trust in political institutions.Newton (2001) rightly observes that like other concepts, political trust has many synonyms that can be used interchangeably with.Notably, expressions like civic-mindedness and participation, citizenship, political interest and involvement, a concern with the public interest/public good, political tolerance, the ability to compromise, and confidence in political institutions may be interpreted as political trust (Newton, 2001).Newton (2001) makes a distinction between political trust and social trust, where the former is learned indirectly through agents such as the media, and the latter is based upon immediate, first-hand experience of others.Schoon and Cheng (2011) define political trust as the confidence people have in their government and institutions.Institutional or "political" trust is defined as trust in societal institutions, as opposed to "generalized" or "social" trust in other people (Lipset and Schneider 1983).
According to Blind (2006), political trust can be directed towards the political system and its organizations as well as the individual political incumbents.Blind (2006) makes a distinction between diffuse and specific political trust, where the former refers to citizens" evaluation of the regime and the political system and the latter refers to assessment of certain political institutions, such as the congress or the local police force.This study examines specific political trust in institutions of presidency, ruling party, parliament, and local council authorities.But first we deal with the determinants of institutional trust below.

Determinants of institutional trust
As it has been mentioned, trust-building is critical for democratic consolidation and ensures that those who have been entrusted with the responsibility to govern do not abuse trust conferred on them by citizens.But across many countries of the developed and developing world, research shows a decline in institutional trust (Inglehart, 2007;Putnam, 2000).Dalton (2005) observes that during the last third of the twentieth century, public trust in government and political institutions eroded in almost all advanced industrial democracies, and in America the decline had to do with political scandals of the 1960s and 1970s.In many advanced democracies, according to Putnam and Goss (2002), there have been changes in the performance of democratic institutions especially the weakening of parties and the decline in public confidence in government.Many governments in industrialized countries have had to deal with harsh economic conditions leading to rise in unemployment and poor delivery of services which ultimately resulted in loss of public confidence.
Africa is not an exception to this trend, as Molomo (2006) points out that "the decline in confidence in the integrity of political institutions and politicians does not emerge in a social vacuum; it is a result of trying social and economic realities in Africa."Armah-Attoh et al. (2007) makes a similar point that in Africa, "political, performance and economic factors (that is, corruption, unfavourable social policy performance, and unfavourable assessments of the general economic and personal living conditions) are the main drivers of institutional trust ratings." A number of studies have actually found that institutional performance is closely connected to trust in institutions.Bratton et al. (2005) observe that, "where government is associated with economic growth, there is more trust in political institutions, because growth implies effective government."Accordingly, people make rational evaluations of how institutions perform and this impact on their trust in such institutions.In developed nations, trust is often low when citizens feel that their governments do not take care of their needs (Blind, 2006).
In the same vein, Miller (1974) concludes that the perception a government has for its citizens that does not function well is associated with distrust.On this basis, the first hypothesis is derived thus: there is a relationship between people"s perceptions of government performance and their evaluation of institutional trust.Citizens who negatively assess government performance are less likely to trust political institutions than those who positively assess government performance.
Theories of social capital found that there is a relationship between political trust and civic engagement or participation in voluntary organizations, even though there is disagreement among scholars on this.According to Van der Meer (2003), most authors using the social capital concept assume that civic engagement and generalised trust influence each other, and that jointly influence the functioning of democracy and therefore trust in political institutions.The concept of social capital has attracted a lot of attention and scholarly interest since the publication of Robert Putnam"s seminal book on Making Democracy Work in which he compared the performance of regional governments in Italy.Putnam (1993) and Mishler and Rose (1999) emphasizes the importance of citizens" embeddedness in a civic community which he defines as dense horizontal networks of associations, which also are typically linked to social structure.Putnam (1993) defines social capital as features of social organization such as trust, norms and networks that can improve the efficiency of society by facilitating co-ordinated actions.According to Putnam (1993), he found that Northern Italy is much more efficient thanks to the cooperative and strongly rooted civic culture compared to the South.He concluded that in a civic community, citizens develop attitudes that enhance cooperation as they are "helpful, respectful, and trustful toward one another".
In this regard, voluntary organizations serve an important role of creating bonds of social solidarity that are the basis for civil society and democracy (Newton, 2001).However, be that as it may, across many nations there is weak evidence of institutional trust by membership of voluntary organizations.As Newton (2001) observes, "There is an association between voluntary activity and social and political trust in some countries, but it is not consistent across nations, and not strong in any".The reason is that too often people spend time either at work, school or with families than they do in voluntary organizations and the cause-effect relationship of civic engagement and trust is difficult to establish.The strongest path is probably that people trust first then join organizations (Newton, 2001).According to Van der Meer (2003) just as much as generalised trust and civic engagement have a reciprocal relationship, it can be assumed that the relation between social capital and trust in political institutions is also reciprocal.
The study second hypothesis states that there is a relationship between civic engagement and people"s evaluation of institutional trust: Citizens who are not active in civic organizations are less likely to trust political institutions than those who are active.Studies have found that institutional trust is related to attitudes on satisfaction with democracy.Newman (2001) posits that "political trust is important because democracies are based on institutional mechanisms that are supposed to ensure that politicians behave in a trustworthy manner."More fundamentally, the worry is that if people do not trust political institutions, which suggests a lack of trust in the manner in which democracy works, and if this happens over an extended period of time, then they may be "disillusioned with democracy as an ideal" (Norris, 1999).
In their study of political trust in new and established democracies, Catterberg and Moreno (2005) argue that in many countries, transition to democracy motivated aspirations of civil, political, and economic rights which placed higher standards for evaluating governmental performance with emergence of democracy.Christensen and Laegreid (2005) note that trust in government generally increases according to the level of satisfaction with democracy, importance of politics in life, interest in politics, membership of political parties and affiliation with the left end of the political spectrum.
Therefore, the expectation is that disaffection with democracy reduces trust in political institutions.Bratton and Gymah-Boadi (2016) succinctly state that "in a democracy, for example, citizens ought to be able to reasonably expect that public officials will govern on their behalf.If, however, government officials are perceived to violate the public"s trust, then people will feel justified in withholding their voluntary compliance" Therefore the study third hypothesis is that there a relationship between people"s satisfaction with democracy and their evaluation of institutional trust.Citizens who are dissatisfied with democracy are less likely to trust political institutions than those who are satisfied with democracy.
Moreover, Institutional trust is bound to be volatile in corruption ridden political systems.Where people perceive corruption in politics, then their perceptions of institutional trust are adversely affected (Job, 2005).Maladministration fosters mistrust among citizens as well as doubts as to the effective enjoyment of legally sanctioned rights (Della Porta and Vannucci, 1997).High levels of corruption undermine both interpersonal and government trust and this has an effect of preventing collective action and the development of civic behavior (Mishler and Rose, 2001).According to Anderson and Tverdova (2003), citizens of countries with high levels of corruption place less value on political institutions and are less confident in their political system.In the same vein, Uslaner (2003) argues that the most corrupt countries have the least trusting citizens.Accordingly, the fourth hypothesis is that citizens who perceive political institutions to be involved in corruption are more likely to mistrust political institutions than those who do not perceive political institutions to be involved in corruption.
Political trust may vary according to certain demographic variables, notably age, education and gender.According to Mishler and Rose (2001), "analysis of political trust emerges from micro-level cultural theories that emphasize that socialization into a culturally homogenous society nonetheless allows substantial variation among individuals based on gender, family background, education, and so forth."A number of studies have actually arrived at different conclusions on the effect of education on Seabo and Molefe 39 institutional trust.For instance Anderson and Singer (2008) show that education tends to boost trust while Seglison (2002) found that the effect of education on trust is negative.But others went further to explore the interactive effect of education with government performance and corruption.
In their comparative study of European countries, Hakhverdian and Mayne (2012) concluded that in countries with comparatively high levels of corruption, education reduces political trust whereas in countries with low levels of corruption, education actually boosts political trust.But the effect of corruption on institutional trust varies with educational attainment because citizens with the lowest levels of education are unresponsive to the effects of corruption but as for other citizens, the effects of corruption on political trust tend to increase with education (Hakhverdian and Mayne, 2012).
Because Botswana is rated the least corrupt African country, the expectation is that generally citizens would trust institutions irrespective of the level of education.But more specifically, the fifth hypothesis is that citizens with low levels of education are likely to trust institutions than people with high level of education.

Construction of the variables
The study dependant variable is built based on the following question in the Afrobarometer survey: "How much do you trust each of the following institutions, or haven"t you heard enough about them to say?"The article considers the answers given regarding the political institutions (president, parliament, local government and ruling party).The relationship between trust and corruption, government performance, satisfaction with democracy and civic participation with other covariates is controlled for.First, a set of demographic variables such as age, gender, level of education and location is introduced.Age could be an important explanatory element of trust in political institutions.Older people are expected to exhibit greater institutional trust because they associate political institutions with the ruling Botswana Democratic Party, which enjoys sentimental attachment to the founding leader Sir Seretse Khama.

RESEARCH DESIGN
Sample surveys are the conventional social-science method for obtaining data about individual attitudes and behaviour.Sample surveys can and do ask individuals to report whether they trust political institutions.In this study analyze the sixth-round Afrobarometer survey conducted in Botswana in 2014 as it is the most recent survey with all the indicators required to test the above hypotheses.In this survey, a nationally representative sample of 1200 Batswana was interviewed.The design is therefore cross-sectional.The model of what explains trust in political institutions stipulates that the likelihood of a person doing so is a function of their spatial location, their evaluation of government performance, satisfaction with political system, social inclusion or

RESULTS
To examine the reliability of the questions on corruption and government performance as measuring a latent variable, factor analysis was also conducted, whereas Cronbach"s alpha (αCr > 0.6) was used as a criterion for the reliability of the extracted factors.Factor analysis of the variable (Q53a-j) on corruption resulted in a onedimensional factor solution (FAC1_1).The Kaiser-Meyer-Olkin measure of sampling adequacy is 0.939 whilst the Bartlett"s test of spherity gave a chi-square value of 10330.823with p=0.000.The KMO statistic is close to 1 whilst test of spherity is highly significant.A measure of the reliability of Q53a-j as measuring a latent variable "corruption", gave a Cronbach"s alpha = 0.947 which is highly reliable.Cronbach's alpha determines the internal consistency or average correlation of items in a survey instrument to gauge its reliability.The analysis of variable Q66a-m on government delivery resulted in a two-dimensional factor solution.A measure of the reliability of Q66a-m gave a Cronbach"s alpha =0.904.
Factor 1 (FAC1_2) covers the provision of basic necessities like water, improving basic health services, addressing education needs etc.The second factor (FAC2_2) can be generalised to cover managing the economy like creating jobs, keeping prices down etc.The Kaiser-Meyer-Olkin measure of sampling adequacy is 0.922 whilst the Bartlett"s test of spherity gave a chi-square value of 6262.212 with p=0.000 suggesting that the R-matrix is not an identity matrix.The KMO statistic is close to 1 indicating that factor analysis will yield distinct and reliable factors.
In Table 1, the study conducts a preliminary analysis to ascertain the strength of the relationship (if any) between trust in political institutions (president, parliament, local government council and ruling party) and each of the independent variables.The results of the Pearson"s correlation coefficient show that almost all the independent variables are significantly correlated the dependent variables at 5% level of significance.The study observation is that the variable gender is not significantly correlated with trust in the president, Parliament and local government council.Gender is however significantly correlated with trust in the ruling party.On civic participation, we observe that joining others to raise an issue is also not significantly correlated with trust in the political institutions under consideration.Membership of a religious group is also not significantly correlated with trust in the president, parliament and ruling party.It is however significantly correlated at 5%significance level with local government council.Education and factor 2 (managing the economy) and level of corruption are significantly negatively correlated with the dependent variable.
In Table 2, a chi-square analysis of each of the dependent variables with the independent variables to assess association is conducted.The chi-square test of association between trust in political institutions and the independent variables show significant association for most variables except gender of respondent and attending community meetings.The variable gender is however only significantly associated with trust in the ruling party with p=0.01.

Regression analysis
The theoretical hypotheses set out earlier can be linked in a simple model.To test these hypotheses, the study

Perception of trust in the president
Prior to fitting a logistic model to predict the likelihood to trust the president shows that gender, an evaluation of the significance of the independent variables was carried out.
Table 3 shows that the residual chi-square statistic is 284.771, which is significant at p=0.000 (labelled overall statistics).This statistics tell us that the coefficients for the variables not in the model are significantly different from zero; in other words, that the addition of one or more of these variables to the model will significantly affect its predictive power.
The remainder of the results in Table 3 lists each of the predictors in turn with a value of Roa"s efficient score statistics for each one (column labelled score).In large, samples when the null hypothesis is true, the score statistics is identical to the Wald statistics and the likelihood ratio statistic.It is used at this stage of the analysis because it is computationally less intensive than the Wald statistic.Roa"s score statistic has a specific distribution from which statistical significance can be obtained.In this table, Q97_R(2), Q97_R(4), Gender(1), FAC1_1, Q19B_R(1) do not look likely to be good predictors because their score statistics are non-significant p>0.05, whilst the rest of the predictors have significant score statistic at p<0.01.
The binary logistic regression model of trust in the president in Table 4 shows that education, level of corruption, government handling of important matters (two factors), joining others to raise an issue and satisfaction with democracy are highly significant explanatory variables in predicting the likelihood to trust in the president.The odds ratio for the independent variable education shows that individuals with informal education are 7.457 times more likely to trust in the president than those with no education; 9.112 times for those with primary education; 5.8 times for individuals with secondary education; 3.588 times for those with post-secondary other than university and 2.572 times for those with university education and above.The odds ratios are much higher for individuals with lower level of education hence disproving the hypothesis that more educated are more likely to trust than the less educated.
Individuals who perceive the level of corruption to have stayed the same are significantly more likely to trust the president (odds ratio = 2.045) and also those who think the level has decreased somewhat or a lot (odds ratio = 3.205).The two factors on government performance; FAC1_2 (provision of basic necessities) and FAC2_2 (managing the economy) are highly significant factor in predicting the likelihood to trust the president.Satisfaction with democracy is a highly significant (p=0.000)factor in explaining the likelihood to trust the president.

Perception of trust in parliament
The residual chi-square statistic (labelled overall statistics) is 168.077, is significant at p=0.000 showing that the coefficients for the variables not in the model are significantly different from zero and therefore, the addition of one or more of these variables to the model will significantly affect its predictive power.
The logistic model of trust in parliament in   The odds ratio for the variable "level of corruption has decreased somewhat or a lot" is 1.829 showing that the probability for an individual to trust parliament is increased compared to someone who said level of corruption has increased somewhat or a lot.Membership of voluntary association increases the likelihood of an individual to trust parliament with an odds ratio of 1.353.The first factor on government performance increases the likelihood of an individual to trust parliament 1.165 times more.

Perceptions of trust in the local government council
Variables that are found to be not significant prior to fitting a logistics model of trust in the local government council are Q97_R(2), Q97_R(3), Q97_R(4), gender (1), age (1), Q54_R(2) and Q19A_R(1).The remaining variables are all significant.
In Table 7, the results show that residual chi-square statistic (labelled Overall Statistics) is 120.919, and is significant at p=0.000 showing that the coefficients for the variables not in the model are significantly different from zero and therefore, the addition of one or more of these variables to the model will significantly affect its predictive power.
The results in Table 8 show that the different levels of education are not significant factors in predicting the likelihood to trust in local government council except informal education which is significant with p=0.01.Individuals who perceive the level of corruption to have decreased somewhat or a lot does not show significant relationship with likelihood of trust in local government council (p>0.05) as well as corruption index.The other non-significant factors (p>0.05) are Q20B_R(1) and age.The second factor on government performance (FAC2_2) and satisfaction with democracy are highly significant with p=<0.001.
The odds ratios show individuals who have informal level of education are more likely (2.561 times) to trust the local government council than somebody with no education.Individuals who perceive the level of corruption to have decreased somewhat or a lot are more likely to trust local government council (odds ratio = 1.372).

Perceptions of trust in the ruling party
The results on Table 9 show the significance level of the variables prior to fitting a logistic model.In Table 9, the results show that residual chi-square statistic (labelled overall statistics) is 262.731, is significant at p=0.000 showing that the coefficients for the variables not in the model are significantly different from zero and therefore, the addition of one or more of these variables to the model will significantly affect its predictive power.The observation made is that Q97_R(4), gender(1), AGE(1), Q54_R(2), FAC1_1, Q19A_R(1) and Q19B_R(1) do not look likely to be good predictors because their score statistics are non-significant p>0.05, whilst the rest of the predictors have significant score statistic at p<0.01.
The results in Table 10 show that the higher levels of education are not significant in predicting the likelihood to trust the ruling party compared to someone with no education.Social inclusion (demographic) variables of location and gender are not significant as well.The logistic regression on trust in the ruling party in Table 10 shows that individuals with informal education are significantly more likely (6.733 times more likely) to trust the ruling party than an individual with no education.For individuals with primary education, the likelihood is increased at 27.201 times more than an individual with no education.
At secondary school level, the likelihood is at 2.763 times an individual with no education.Perceptions of level of corruption are also a significant predictor: individuals who perceive level of corruption to have decreased

DISCUSSION
The findings confirm the fifth hypothesis that citizens with lower levels of education tend to trust political institutions.Specifically, less educated Batswana trust the president, ruling party and local government authorities.This may be due to the visibility of the president, and by extension the ruling party as well as the proximity of local government councils to such people, who in most cases reside in rural areas.The president"s walkabouts in villages and sitting around the fire with elderly people to share dinner has endeared him to the rural people.However, the level of education is not a predictor of the likelihood to trust parliament partly because parliament is often viewed to be detached from the electorates and confined to the capital city.Batswana often decry that members of parliament are only visible during elections campaigns and after being voted they forget about the electorates.
In terms of government performance, provision of basic necessities (FAC1_2) is a significant factor in predicting the likelihood to trust the president, parliament and the ruling party, but not for local government council.This confirms the first hypothesis that citizens who are content with government performance trust institutions.It appears that Batswana attach the provision of such services to the president, parliament and ruling party since central government has since centralised most of the basic services including health, water and sanitation and basic education in 2008.In particular, services such as primary health care and primary education used to fall under the purview of councils but have since been centralised.It is not surprising therefore, that Batswana"s perceptions of trust in councils are not a function of provision of basic necessities.Managing the economy (FAC2_2) is a significant predictor of the likelihood to trust political institutions.
The results also indicate that individuals who perceive the level of corruption to have stayed the same/decreased somewhat or a lot are more likely to trust the president, parliament, local government and ruling party.The inverse relationship between perceptions of level of corruption and trust could explain the declining level of trust in political institutions since 2008.Satisfaction with democracy is a highly significant factor that shapes Batswana"s perceptions of trust in political institutions.With this in mind, the period between 2008 and 2014 has witnessed a decline in Botswana"s democracy particularly media freedom, allegations of extra judicial killings and violation of minority rights.
In terms of civic engagement, the hypothesis finds support when people get together to raise an issue than their membership to religious groups and voluntary association.This is not surprising because communities raise issues that affect them with political leadership in most instances through the Kgotla platform.Political leaders especially of the BDP prefer to address issues raised by communities in the Kgotla meetings, due to the cultural importance of the institution which grants legitimacy to decisions reached.But besides, political leaders use the kgotla platform for political convenience.
The data confirms the literature that the relationship between membership of voluntary organizations and political trust is not strong.Although Batswana are religious and affiliated to various church denominations, religion still remains a private matter and it is not influential in politics.In short, Botswana is a secular state, where the state remains neutral in matters of religion.The implications of these results are that more work on this area has to be conducted to even investigate reasons that account for a decline in institutional trust.To this end, there still remains a yawning gap in the literature on this subject.

Conclusion
This study has examined factors that influence trust in political institutions.It has revealed that although Botswana is widely acclaimed for performing well in good governance relative to other African countries, the citizens are increasingly becoming weary of political institutions and losing trust in them.This, the paper establishes, has got to do with the disaffection with democracy and government performance in delivery of essential services.The level of education is important because people of lower education tend to trust political institutions particularly the president, the ruling party and local government council.But this is not true for parliament which is viewed as detached from the electorates.The paper has argued that Batswana are critical of their institutions irrespective of their seeming lack of civic engagement.The implication of this study results is that political institutions may have to deal with critical citizens who would demand answers for poor performance more than ever before.

Table 1 .
Pearson"s correlation coefficient between dependent and independent variables.
** Correlation is significant at the 0.01 level (2-tailed); *Correlation is significant at the 0.005 level (2-tailed).exclusion, and corruption.The level of analysis will be individual Batswana who are of voting age (18 years+).

Table 2 .
Chi-square test of association.

Table 3 .
Significance of variables not in the model.

Table 4 .
Logistic regression model of perceptions of trust in the president.
Table6shows that the first factor on government delivery (FAC1_2) is not

Table 5 .
Significance of variables not in the model.

Table 6 .
Logistic regression model of perceptions of trust in parliament.

Table 7 .
Significance of variables not in the model.

Table 8 .
Logistic regression model of perceptions of trust in local government council.