Rural Household's Intention to Use Microfinance in Tanzania. (2024)

I. INTRODUCTION

Microfinance, which include micro-credit, savings andmicro-insurance (Consultative Group to Assist the Poor, 2012) is aneconomic development approach used by the government to assisthouseholds who are not qualified for conventional financial schemes(Ashraf and Ibrahim, 2014; Yuge, 2011). Microfinance has been asuccessful tool in helping low-income households to accumulate assets,boosting their incomes and eventually improving their standard ofliving. Despite of the implementation of various financial promotionalcampaigns activated by private and public sectors, the adoption rate ofmicrofinance among the households was low in developing countries, forexample 22% in Tanzania and 24% in Sub-Saharan Africa compared to 50% inthe worldwide (Demirguc-Kunt and Klapper, 2012; Bank of Tanzania, 2013).

In Tanzania, most of the households who are not qualified forconventional banking services live in rural areas (National Bureau ofStatistics 2013). Despite of the diverse government efforts tofacilitate provision of financial services in the rural areas, yet thepoverty level in Tanzania is high (28.2%) compared to other Africancountries like Uganda and Botswana (FinScope, 2013). Moreover about 90%of the poor people live in the rural areas. Therefore, microfinancecould enable rural households to boost their income through improvedagricultural output by employing modern farming techniques. In Tanzaniaagriculture is considered to be the backbone of the economy because itcontributes large proportion of the Gross Domestic Product (GDP),employment, raw materials for industries and national food security(Ministry of Agriculture Food Security and Cooperatives 2015).

Formal microfinance services are provided by licensed commercialbanks, non-bank financial institutions, savings and credit cooperativesocieties (SACCO) and financial non-government organizations (NGO)(Triodos Facet 2007). To enable more targeted users to meet the minimumrequirements for the financial services, the service providers encouragethe potential users to apply microfinance services such as micro-loansin groups, rather than individually. Low adoption of microfinance couldbe plausibly caused by the following reasons: the fear of losingpersonal assets if they fail to settle their own or other groupmember's outstanding loans (termed as perceived barrier) andthereby increases their unfavourable attitude towards the use ofmicrofinance which in turn reduces their intention to adopt microfinance(Bank of Tanzania, 2014; Mahlanza, 2015).

Secondly, limited knowledge of the benefits that could be acquiredfrom the usage of microfinance could cause the rural households todevelop unfavourable attitude towards microfinance and this couldeventually decrease their intention to adopt microfinance too (FSDT,2014). Thirdly, social pressure that need to be borne by households as aresult of negative opinions given by people who are important to them(termed as subjective norm) may affect the household's intention toadopt microfinance services (Chogo and Sedoyeka, 2015). The lack ofbasic financial knowledge on how to utilise the microfinance services(termed as perceived behavioural control) to increase their farming orbusiness outputs could reduce the household's adoption intention ofmicrofinance as well (Bank of Tanzania, 2014). Therefore, this studyintends to achieve the following objectives in order to address theproblems above.

A. Research Objectives

1) To examine the direct effects that can be generated by perceivedbarriers, perceived benefits, attitude, subjective norm, and perceivedbehavioural control on households' intention to use microfinance inTanzania; and

2) To investigate the mediation effects of perceived benefits andperceived barriers on households' intention to use microfinancethrough attitude.

II. LITERATURE REVIEW

Microfinance studies in Tanzania tended to focus on solving issuesrelated to the impacts that can be generated by microfinance to studiedcommunity, loan repayment behaviours, savings behaviour, and factorsthat could limit the growth of microfinance institutions (Mukama et al.,2005; Kato and Kratzer, 2013; Kessy and Urio, 2006; Kipesha, 2013). Mostof the previous studies have confirmed the ability of microfinance toenhance agricultural productivity (Girabi et al., 2013). However,because of various factors that include high interest rates and lack ofunderstanding of microfinance benefits deterred some of the Tanzaniansrural households' intention to use microfinance services. Ahlen(2012); Mohamed and Ahmed (2015); Morduch and Haley (2002) support thepositive impact of microfinance on improving households ability to meetconsumption needs, enhance investments, improve living standards andreduction of poverty. Mwatsika (2015) claims that rural communitiescould be aware of the existence of the formal financial institutions yetthey do not utilize services. The non-utilization could probably resultfrom rural community's low level of education, excessive charges bythe MFIs and stringent requirements for collateral.

Notwithstanding diverse studies that were carried out aboutmicrofinance in Tanzania, however limited studies investigatedbehavioural factors that could influence household's intention touse microfinance. In order to address the behavioral problems of thisstudy the theory of planned behaviour (TPB) was adapted. TPB can provideexplanation on how individual's perceptions influence their actions(Ajzen 1991). The understanding of the behavioral factors that influencehouseholds' intention to use microfinance services provides usefulinput for policy makers and practitioners to promote the use ofmicrofinance to increase agricultural production. For instance theincreased agricultural production facilitate the achievement of thecountry's vision 2025 of being the middle income country likeMalaysia and Indonesia (The United Republic of Tanzania President'sOffice Planning Commission, 2009).

Various behavioural theories can be used to address currentresearch problems such as theory of reasoned action (TRA), theory ofplanned behaviour (TPB), and technology acceptance model (TAM) (Ajzen,1991; Fishbein and Ajzen, 1975; Davis, 1989). TRA explains that anindividual will perform certain act if that person has positiveintention to perform that act and the behavioural intention is driven bythe effects generated by attitude and subjective norms (Fishbein andAjzen, 1975). The theory assumes that respondents' behaviour isunder volitional control and therefore, TRA may not be suitably used asa basic research model if the respondent is constrained by certainfactors such as possessing only limited knowledge, skills, money orcooperation from others. As the result, Ajzen (1991) modified TRA byincluding an additional variable, perceived behaviour control in a newmodel, theory of planned behaviour (TPB).

The TPB model has been used in various research areas that havebeen carried out worldwide such as adoption of electronic banking;performing entrepreneurial business; and using mobile healthcareservices (Krueger Jr. et al., 2000; Shanmugam et al., 2014; Deng et al.,2013). Most of the past studies have confirmed the relationships betweenTPB's constructs (Kisaka, 2014; Abdul Adis et al. 2015,).Therefore, this study is using TPB as the basic theory because thetheory can address the problems related to attitude, subjective norm,and perceived behavioural control. Furthermore, limited studies haveused TPB to predict households' intention to use microfinance in adeveloping country rural setting.

Ajzen (1991) encouraged researchers to modify TPB model if themodification can explain the respondents' behavioral intention morecomprehensively. Thus, this study modified TPB by adding two variables:perceived barriers and perceived benefits to address currentstudy's research problems (see Figure 1). Both perceived barriersand perceived benefits variables had been tested in different researchcontexts such as computer security, reading and writing short messages(SMS) driving behaviour and water saving behaviour (Claar and Johnson,2010; Chuah et al., 2016; Knowlden and Sharma, 2012; Morowatisharifabadet al., 2012). However, limited studies in finance context have testedthe direct and indirect effects that can be created by TPB'sconstructs (attitude, subjective norms and perceived behaviour control),perceived barriers and perceived benefits on behavioural intention.

A. Conceptual Model and Hypotheses

Figure 1 presents the theoretical model that was developed for thisstudy based on the theory of planned behaviour (TPB). The research modelindicates the constructs that influence rural households' intentionto use microfinance services. To examine the direct relationship betweenthe independent constructs (perceived barriers, perceived benefits,attitude, subjective norm, and perceived behavioural control anddependent construct (behavioural intention) five hypotheses weredeveloped (H1, H3, H5, H6, and H7). Furthermore, the mediation effectwas investigated through developing two hypotheses (H8a and 8b).Therefore, seven hypotheses were developed to achieve two researchobjectives of this study. However, it was necessary to develop otherhypotheses (H2 and H4) to complement the mediation analysis, but theyhad no direct impact on the dependent variable.

In current study, perceived barriers represents household'sbelief about the probable negative consequences such as perceivedtangible and intangible costs that a person need to bear if microfinanceis adopted. This variable can generate direct and indirect effects tobehavioural intention. If respondents strongly believe that negativeconsequences may follow up as the result of their engagement inmicrofinance, their behavioural intention could be directly affected.Studies in different research contexts had confirmed the direct andnegative relationship between perceived barriers and behaviouralintention (Cacciotti and Hayton, 2014; Jebarajakirthy and Lobo, 2014;Leung et al, 2012; Shariff, Amran, and Goh, 2012; Shinnar, Giacomin, andJanssen, 2012). In the meantime, indirect effects of perceived barrierscan happen as well. Respondents' attitude could become morenegative if they may need to endure more perceived barriers and thiswill eventually reduce their intention to adopt microfinance (Ifeluninand Elizabeth 2013). This is supported by the studies carried out byArnaut (2015); Jebarajakirthy and Lobo (2014); Malebana (2015).

In summary, we hypothesize that:

H1: Perceived barriers can generate direct negative effect on thehouseholds' intention to adopt microfinance.

H2: Perceived barriers can generate direct negative effect on thehouseholds' attitude.

From H1 and H2, we predict that:

H8a: The relationship between perceived barriers andhouseholds' intention is mediated by households' attitude.

The term perceived benefits is referring to households'perception about the probable positive outcome that can be gained as theresult of microfinance adoption. Past studies suggest that perceivedbenefits could generate direct and positive relationship on behaviouralintention (Lin et al., 2013; Ng et al., 2009; Shanmugam et al., 2014;Tanadi et al., 2015). This implies that individuals' lack ofunderstanding of the benefits which could be generated from the use ofcertain financial products or services would lead to decline of thebehavioral intention. In addition, past studies also suggest that theeffect of perceived benefits on behavioural intention could be mediatedby individuals' attitude toward the studied subject (Liu et al.,2012; Wani and Malik, 2013; Lee, 2009; Shanmugam et al., 2014). Ruralhouseholds' believe about the ability of microfinance to accomplishfinancial needs and availability of variety of financial services inTanzania would develop favourable attitude and subsequently enhancetheir intention to use microfinance. Therefore, this study predictsthat:

H3: Perceived benefits can generate direct positive effect on thehouseholds' intention to adopt microfinance.

H4: Perceived benefits can generate direct positive effect on thehouseholds' attitude.

From H3 and H4, current authors summarize the effects as follows:

H8b: The relationship between perceived benefits and intention toadopt microfinance is mediated by households' attitude.

The attitude shows the degree to which a household has favourableappraisal about their participation in microfinance services. Thefollowing researchers suggest that attitude could generate direct andpositive relationship on behavioural intention: Alam, Janor, Zanariah,Wel, and Ahsan (2012); Putit and Johan (2015); Shah Alam and Sayuti(2011); Tsordia and Papadimitriou (2015). Therefore, households'intention to use microfinance would be higher if they favour theadoption of microfinance.

H5: Households' attitude toward the adoption of microfinancecan generate direct positive effect on their intention to adoptmicrofinance.

Subjective norm reflects the perceived social pressure thathouseholds need to undertake if they are interested to use microfinance.The social pressure could be from family members, friends, siblings orcommunity members who could influence respondents' intention toperform certain behaviour. According to past researchers, if respondentsare willing to align their behaviour according to other people'sopinion, the relationship between subjective norm and behaviouralintention would be positively related (Alsmady, Rahman, and Muhammad,2014; Ashraf and Ibrahim, 2014; Mishra, 2014; Tsordia and Papadimitriou,2015). As the current respondents lack knowledge on how microfinancecould assist them to increase the business outputs, they may need torefer to other people who are important to them before adopting themicrofinance. Furthermore, Tanzanians are practicing collectivistculture. Based on that premise, this study predicts that:

H6: Subjective norm can generate direct positive effect onhouseholds' intention to adopt microfinance.

Perceived behaviour control refers to households' perceptionof the feasibility to use microfinance given the presence or absence ofindispensable resources and opportunities. This implies that respondentswould have higher intention to use microfinance if they have requisiteresources such as knowledge, skills, or cooperation from other people orpublic or private agencies. The relationship between perceived behaviourcontrol and behavioral intention had been confirmed in differentresearch contexts such as Islamic banking services, Internet stockbroking, and commercial zakat service (Amin et al., 2014; Heikal andKhaddafi, 2014; Gopi and Ramayah, 2007; Rezaie and Abadi, 2012; Alam etal., 2012). Hence, the current study hypothesizes that households'behavioural intention will increase when they have the necessaryresources and/or could become more knowledgeable about microfinance.

H7: Perceived behaviour control can generate direct positive effecton households' intention to adopt microfinance.

III. RESEARCH METHODOLOGY

Majority of Tanzanians (70.9%) live in rural areas and most of themare engaged in agricultural activities (National Bureau of Statistics2013). Therefore, this study is targeting the rural households, aged 18years and above that own an average 0.9 to 3 hectares of land (Rugumamu,2014; Wolter, 2008) and have not adopted microfinance scheme.Respondents were selected from five geographical zones: coastal,northern, southern highland, lake zone and central zone because majorityof the rural households lived in the areas (National Bureau ofStatistics, 2014).

This study adopted quantitative research approach because itpermits testing of the proposed research hypotheses through inferentialstatistics and the generalization of the study results among the ruralhouseholds. The total population of rural households in Tanzania thatengage in farming activities and aged 18 years or more was 8,788,143(National Bureau of Statistics, 2014). The minimum sample size of 400respondents was determined through the Yamane (1967) formula where itsassumptions were met in this study that include random sampling andhom*ogeneity of population elements.

Stratified and multistage probability sampling techniques were usedbecause the population in each zone could be equitably selected as thestudy's respondents. In the first stage of sampling process,population is divided into five geographical zones in Tanzania mainland.In the subsequent stages, multistage sampling technique is used toselect the representative regions, districts, wards and villages locatedin each zone. In the final stage, 600 respondents were selected fromfive villages by using simple random technique.

Questionnaires were administered by using drop-off and pick-up(DOPU) method. The method can provide higher response rate because thequestionnaires were physically distributed to and collected fromrespondents on the same or following day. Other distribution methodssuch as postage mails or emails could not be suitable for this studybecause the infrastructure and communication system in rural areas werenot well developed. Five hundred thirty six questionnaires werecollected and 47 questionnaires were voided due to the missing data.Nevertheless, the amount of useable answered questionnaire is sufficientfor statistical analysis tests (Schlomer et al., 2010; Chandio, 2011).

Each questionnaire was prepared in two languages: English andSwahili (the national language in Tanzania). Each studiedvariable's item was measured by using 5-point Likert scale; rangingfrom strongly disagree (1) to strongly agree (5). Pretest was carriedout to ensure that the statement used to measure each item in currentstudy's questionnaire could reflect the original intendedmeasurement. According to Presser et al. (2004), sample size of between20 to 50 respondents is adequate for pretesting. Therefore, fiftyacademic and financial industry experts were employed to suggestimprovement on each item statement's lengthiness, formatting, andwording.

Following the pretest, minor amendments were made and amendedquestionnaires were distributed for the main study. The mainstudy's quantitative data were analysed by using two stagesapproach (Anderson and Gerbing, 1988). First, confirmatory factoranalysis (CFA) was used to validate measurement model by assessing theunidimensionality, validity and reliability of latent variables. Then,structural model was built up by using IBMAMOS version 21.0 to test andconfirm the hypotheses, subsequent the validation of the measurementmodel.

Furthermore, bias corrected bootstrap method was employed to assessthe mediation effects (H8a and H8b) because of its strong statisticalpower and it does not impose strict assumptions about normality of thedata distribution (Preacher and Hayes, 2008; Fritz et al., 2012).According to Schleider, Patel, Krumholz, Chorpita, and Weisz (2015),mediation effect would occur if the lower and upper boundary determinedat 95% confidence interval do not include zero.

IV. RESULTS AND DISCUSSIONS

A. Respondents' Demographic Profiles

The respondents profile indicates about 98% of the participants whowere within the productive age of 18 to 64 years. In terms of genderdistribution, the proportion of female respondents (59%) was slightlyhigher than males (41%). This demonstrates that rural populationconsists of more females compared to men (Food and AgriculturalOrganization, 2014). The results indicate the majority of therespondents were married (68%) and few of them were single (32%).Moreover, about 89% of the participants had non-tertiary education leveland 11% had tertiary education. The results reflect the highest level ofeducation achieved by the majority of the individuals in the rural area.

B. Structural Equation Modelling

In order to achieve the objectives of the current study, thestructural equation modelling (SEM) with two stages approach was used.

1. Assessment of measurement model

Table 1 indicates that the fit indices for initial measurementmodel were below the acceptable threshold (Schumacker and Loamx, 2010;Kline, 2011). Poor model fit was detected as a result of the low factorloading scores (below 0.5) shown by five measurement items (Awang, 2015;Fornell and Larcker, 1981).

To improve the unidimensionality and model fit, five iterationswere run to improve the model fitness by deleting the low factor loadingitems. The unidimensionality is achieved when all measurement itemsachieve the acceptable factor loading (Awang 2015). The deletion processstarted with an item that had the lowest factor loading, then re-run themodel until the unidimensionality was achieved. In summary, the fiveitems: each from perceived behaviour control, behavioural intention andperceived barrier and two items from perceived benefits were deletedfrom the final measurement model. Following the elimination of the fiveitems that had poor factor loading, the unidimensionality was achieved(loading scores exceed 0.5) and the modified measurement model haddesirable goodness of fit to the data (Figure 2).

Construct validity is achieved if the studied measurement itemscould reflect the respective latent variable accurately. In confirmingthe construct validity, convergent and discriminant validity tests werecarried out. The convergent validity was tested by computing the scoresof average variance extracted (AVE) and composite reliability (CR). Theresults shown in Table 2 indicates that the respondents had providedconsistent responses for all measurement items used to measure eachvariable as the AVE and CR scores were higher than the threshold valuesof 0.5 and 0.7 respectively (Bagozzi et al., 1992; Hair et al., 2014;Alallwan et al., 2017).

Discriminant validity test was carried out to confirm whether themeasurement items of each studied variable differs from the items usedto measure other variables. Table 3 shows that the square root of AVEfor each variable is larger than inter-construct correlation scores withother corresponding constructs (Kline, 2011; Fornell and Larcker, 1981).Thus, the results demonstrate that discriminant validity is achieved.

2. Model fit indices

The structural model yielded after achieving acceptable model fitstatus is shown on Figure 3. The examined model fit indices were withinthe required threshold as follows: GFI= 0.914; SRMR= 0.05; RMSEA =0.059; CFI= 0.923; NNFI= 0.908; and Chi-square/df = 2.702. Thus, theresults imply that current theoretical model fits well to the observedsample data.

3. Confirmation of direct effect hypotheses

Table 4 presents the results of seven direct hypotheses tested inthis study. The findings show that six hypotheses (H1, H3, H4, H5, H6,and H7) are supported and hypothesis H2 is not supported. To elaborate,perceived barriers and behavioral intention (H1) are confirmed to bestatistically significant negative related. In other words, when theperceived barrier is high, households' intention to adoptmicrofinance will decrease. The finding is consistent with the followingpast studies', Jebarajakirthy and Lobo (2014); Liu et al. (2012).

The confirmation of H3 shows that perceived benefits and behavioralintention is statistically positive related and the results supportstudies carried out by Lee (2009) and Shanmugam; Savarimuthu, and Wen(2014). Therefore, if the benefits of using microfinance are welldisseminated, respondents' intention to adopt microfinance willincrease. The research findings support H4 that perceived benefitssignificantly influence the households' attitude towardsmicrofinance and the result is consistent with studies carried out byLin, Hsu, and Chen (2013); Shanmugam et al. (2014); Liu et al. (2012).If the households' believe that the use of microfinance can improvetheir business's outputs, they will favour the use of microfinance.

Table 4 also shows that attitude can affect respondents'behavioural intention, reflected by the support of hypothesis H5.Consistent with the past studies' results (Heikal and Khaddafi,2014; Sayid and Echchabi, 2013), households' intention to adoptmicrofinance will increase if favourable attitude towards themicrofinance scheme can be formed. Confirmation of hypothesis H6 alsosupports that subjective norm has positive significant influence onrespondents' behavioural intention. The result is consistent topast studies' findings (Karlan and Zinman, 2011; Alqasa, 2014; Aminet al., 2014; Putit and Johan, 2015). The result implies that socialpressure plays a significant role on influencing households'behavioural intention. According to H7, perceived behaviour controlcould positively affect the change of respondents' intention toadopt microfinance. The finding supports studies carried out by Alam etal. (2012); Phan and Zhou (2014); Prabandari and Sholihah (2014);Safeena et al. (2013) - households' intention to adopt microfinancecould increase if their knowledge about microfinance increases.

4. Confirmation of indirect effect hypotheses

Table 5 shows that the mediation effect of attitude on therelationship between perceived benefits and households' intentionto adopt microfinance is statistically significant at confidenceinterval of 95%. Therefore, hypothesis (H8b) was supported andconsistent with the studies of Lin et al. (2013) and Shanmugam et al.(2014). On the other hand, hypothesis (H8a) was not supported. Plausiblythis is because the policy makers have been active in disseminating thebenefits of using microfinance and less active in alerting the potentialusers of the barriers that they need to undertake.

V. DISCUSSION OF MAIN RESULTS, IMPLICATIONS AND CONCLUSION

The results show that perceived benefits play a significant role inaffecting households' intention to adopt microfinance directly andindirectly. As rural areas' infrastructure and communicationsystems were not well developed, policy makers should ensure thatinformation related to the benefits of microfinance could reach therespondents. In this context, it may be wise to involve people who areimportant to the respondents. The support of the positive effectgenerated by subjective norm on behavioural intention (reflected by theconfirmation of H6) shows that households would adopt microfinance ifthey receive positive comments about microfinance from their familymembers and social community. For example, awareness campaign about themicrofinance benefits can be conducted through village assemblies anddisseminated by trained agents or merchants. On top of affecting thehouseholds' behavioural intention directly (H3), such promotionmethod could increase the respondents' favourable attitude towardsmicrofinance (H4) as well. In other words, attitude can mediate theeffects of perceived benefits on behavioural intention (H8b).

On the other hand, perceived barriers can affect the change ofbehavioural intention directly (H1) and may not be able to change therespondents' attitude (H2). Possibly, this is because thehouseholds are alert that if certain barriers exist, their intention toadopt microfinance will directly decrease. However, as the barriers thatthey need to face upon adopting the microfinance are not well published,they may not be able to respond on how barriers could affect theirattitude. Therefore, households could not respond significantly on themediation effect of attitude on the relationship between perceivedbarriers and behavioural intention (see the result of H8a in Table 5).

To enhance the direct impact, policy makers should reduce the:waiting time to access the microfinance services; loan's interestrates; and the possible fear of losing cash or personal assets pledgedas collateral. Nurturing positive attitude toward microfinance should beexercised because the support of H5 shows that the relationship betweenattitude and behavioural intention is statistically significant. Afterreducing the barriers, policy makers should publicise it so thathouseholds would become favourable towards microfinance and eventuallyincreasing their adoption intention.

The organised public financial training should be carried outthrough mass media (such as newspapers, radios, and television),organised seminars, credit-linked training programs and villageassembles. The financial trainings would equip the households withknowledge on how to utilise the financial services that could eventuallyincrease their intention to adopt microfinance, which is supported bythe significant result of H7.

In summary, perceived benefits, perceived barriers, attitude,subjective norm, and perceived behavioural control could playsignificant roles in affecting household's behavioural intention.To form favourable attitude towards the adoption of microfinance, policymakers should disseminate the benefits of the financial services moreeffectively and be more transparent in disclosing the possible barriersthat the targeted users may need to bear.

Generally, this study has enriched the theory of planned behaviourin finance by incorporating additional variables: perceived barriers andperceived benefits, and measuring the mediating effects that could beplayed by attitude. Moreover, the mediation effects were tested by usingbias corrected bootstrap method which is a sophisticated technique thatproduced more accurate results about the true mediation effects relativeto normal theory approaches used by most of the past studies (Sobel,1982; Hayes, 2009; Baron and Kenny, 1986).

VI. RESEARCH LIMITATIONS AND SUGGESTIONS FOR FUTURE RESEARCH

The conclusion drawn in this study is based on cross sectionaldata, in which data was collected over a single period. As a result, itmight limit the application of this study findings in future if certainfactors have changed such as respondents' level of education.Therefore, continuous studies should be carried out constantly. Thisstudy focused on the effect that can be generated by behavioural factorson households' intention to adopt microfinance in a rural setting.Hence, this could limit the generalization of the present study findingsamong the households in urban setting because their perception could bedifferent. Current authors recommend future researcher to replicate thepresent research model for different settings.

REFERENCES

Abdul, A., G. Ing, Z. Osman, I. Razli, Y. Pang, S. Sondoh, and M.Abdul, 2015, "Mediation Role of Attitude towards Product Placementin Social Media." Journal of Sustainable Development, 8(3), 79-88.

Ahlen, M., 2012, Rural Member-Based Microfinance Institutions -Afield study assessing the impacts of SACCOS and VICOBA in Babatidistrict, Tanzania. Soderton University. Available at:www.diva-portal.org/smash/get/diva2:544512/FULLTEXT01.pdf.

Ajzen, I., 1991, "The theory of planned behavior."Orgnizational Behavior and Human Decision Processes, 50, 179-211.

Alallwan, A., Y. Dwivedi, and N. Rana, 2017, "InternationalJournal of Information Management Factors Influencing Adoption of MobileBanking by Jordanian Bank Customers: Extending UTAUT2 with Trust."International Journal of Information Management, 37, 99-110.

Alam, S., H. Janor, Zanariah, C. Wel, and M. Ahsan, 2012, "IsReligiosity an Important Factor in Influencing the Intention toUndertake Islamic Home Financing in Klang Valley?" World AppliedSciences Journal, 19(7), pp.1030-1041.

Alqasa, K.M., 2014, "The Impact of Students' Attitude andSubjective Norm on the Behavioural Intention to Use Services of BankingSystem." Int. J. Business Information Systems, 15(1), 105-122.

Alsmady, A.A., R.A. Rahman and A.D. Muhammad, 2014, "EthicalResponsibility and Applicability of Islamic Micro-Investment Model(IMIM) in Jordan." Middle-East Journal of Scientific Research,22(5), 743-751.

Amin, H., A.R.A. Rahma and D.A. Razak, 2014, "Willingness toBe a Partner in Musharakah Mutanaqisah Home Financing : EmpiricalInvestigation of Psychological Factors." Jurnal Pengurusan, 40,69-81.

Anderson, J.C. and D.W. Gerbing, 1988, "Structural EquationModeling in Practice: A Review and Recommended Two-Step Approach."Psychological bulletin, 103(3), 411-423.

Arnaut, D., 2015, "Analysis of Fear as Barrier toEntrepreneurship." International Journal of Economics, Commerce andManagement, III(5), 854-861.

Ashraf, M.A. and Y.B. Ibrahim, 2014, "Poverty Alleviation andIdentifying the Barriers to the Rural Poor Participation in MFIs: A CaseStudy in Bangladesh 1 Mohammad A. Ashraf 2 and Yusnidah B. Ibrahim3." Journal of Economics Cooperation and Development, 35(3),99-132.

Awang, Z., 2015, SEM Made Simple: A Gentle Approach to LearningStructural Equation Modeling, Selangor: MPWS Rich Publication Sdn. Bhd.

Bagozzi, R.P., H. Baumgartner and Y. Yi, 1992, "State VersusAction Orientation on the TRA Action: An Application of CouponUsage." Journal of Consumer Research, 18(4), 505-518.

Bank of Tanzania, 2013, "Dar es salaam: BIS CentralBankers' Speeches." Addressing the challenges of financialinclusion in Tanzania. Available at:http://www.bis.org/review/r131217f.pdf.

Bank of Tanzania, 2014, Directorate of Banking Supervision AnnualReport, Dar es salaam. Available at:http://www.bot.go.tz/BankingSupervision/Reports/DBSANNUAL REPOR2014.pdf.

Baron, R.M. and D. Kenny, 1986, "The Moderator-MediatorVariable Distinction in Social The Moderator-Mediator VariableDistinction in Social Psychological Research: Conceptual, Strategic, andStatistical Considerations." Journal of Personality and SocialPsychology, 51(6), 1173-1182.

Cacciotti, G. and J.C. Hayton, 2014, "Fear of Failure andEntrepreneurship: A Review and Direction for Future Research."Enterprise Research Centre, (24), 1-62.

Chandio, F., 2011. Studying Acceptance of Online BankingInformation System: A Structural Equation Model. PhD Thesis, London, UK.

Chogo, P.J. and E. Sedoyeka, 2015, "Exploring FactorsAffecting Mobile Money Adoption in Tanzania." International Journalof Computing and ICT Research, 8(2), 53-64.

Chuah, F., H. Ting, E. De Run, and J. Cheah, 2016,"Reconsidering What Entrepreneurial Intention Implies: The Evidencefrom Malaysian University Students." International Journal ofBusiness and Social Science, 7(9), 85-98.

Claar, C.L. and J. Johnson, 2010, "Analyzing the Adoption ofComputer Security Utilizing the Health Belief Model." Issues inInformation Systems, 11(1), 286-291.

Consultative Group to Assist the Poor, 2012, A Guide to -Regulation and Supervision of - Microfinance; Consensus Guidlines,Washington D.C. Available at:www.cgap.org/sites/default/files/Consensus-Guideline-A-Guide-to-Regulation-and-Supervision-of-Microfinance-Oct-2012_0.pdf.

Davis, F.D., 1989, "Perceived Ease of Use, and User Acceptanceof Information Technology." MIS Quarterly, 13(3), 319-340.

Demirguc-Kunt, A. and L. Klapper, 2012, Financial inclusion inAfrica: an overview, World Bank Policy Research.

Deng, Z., X. Mo and S. Liu, 2013, "Comparison of theMiddle-Aged and Older Users' Adoption of Mobile Health Services inChina." International Journal of Medical Informatics, 83(3),210-224.

FinScope, 2013, A Survey of the Demand for financial Services inTanzania, Dar es salaam. Available at:http://www.fsdt.or.tz/wp-content/uploads/2016/05/FinScope-Brochure-2013-Summary-2.pdf.

Fishbein, M. and I. Ajzen, 1975, Belief, Attitude, Intention andBehaviour: An Introduction to theory and Reseach, Amherst: Addison-Wesley. Available at: philpapers.org/rec/FISBAI.

Food and Agricultural Organization, 2014, Gender Inequalities inRural Employment in Tanzania Mainland Gender Inequalities in RuralEmployment in Tanzania Mainland: An Overview, Rome. Available at:http://www.fao.org/3/a-i4083e.pdf.

Fornell, C. and F.D. Larcker, 1981, "Evaluating StructuralEquation Models with Unobservable Variables and Measurement Error."Journal of Marketing Research, 18(1), 39-50.

Fritz, M.S., A.B. Taylor and D.P. Mackinnon, 2012,"Explanation of Two Anomalous Results in Statistical MediationAnalysis." Multivariate behavioral research, 47(1), 61-87.

FSDT, 2014, Promoting financial inclusion in Tanzania, Dar essalaam. Available at:http://www.fsdt.or.tz/wp-content/uploads/2016/05/FSDT-Our-Work-2013-2014-Full-Report.pdf.

Girabi, F., G. Mwakaje and A. Elishadai, 2013, "Impact OfMicrofinance On Smallholder Farm Productivity In Tanzania: The Case OfIramba District." Asian Economic and Financial Review, 3(2),227-242.

Gopi, M. and T. Ramayah, 2007, "Applicability of Theory ofPlanned Behavior in Predicting Intention to Trade Online."International Journal of Emerging Markets, 2(4), 348-360.

Hair, J., M.L.D. da S. Gabriel, and V. Patel, 2014, "AMOSCovariance-Based Structural Equation Modeling (CB-SEM): Guidelines onits Application as a Marketing Research Tool." Revista Brasileirade Marketing, 13(2), 44-55.

Hayes, A.F., 2009, "Beyond Baron and Kenny: StatisticalMediation Analysis in the New Millennium." Communicationmonographs, 76(4), 408-420.

Heikal, M. and M. Khaddafi, 2014, "The Intention to Pay ZakatCommercial: An Application of Revised Theory of Planned Behavior."Journal of Economics and Behavioral Studies, 6(9), 727-734.

Ifelunin, C. and W. Elizabeth, 2013, "Constraints to WomenEntrepreneurs' Access to Microfinance in Bayelsa state,Nigeria." Research Journal of Finance and Accounting, 4(6), 6-13.

Jebarajakirthy, C. and A.C. Lobo, 2014, "War Affected Youth asConsumers of Microcredit: An Application and Extension of the Theory ofPlanned Behaviour." Journal of Retailing and Consumer Services,21(3), 239-248.

Karlan, D. and J. Zinman, 2011, "Microcredit in Theory andPractice: Using Randomized Credit Scoring for Impact Evaluation."Science, 332(6035), 1278-1284.

Kato, M.P. and J. Kratzer, 2013, "Empowering Women throughMicrofinance : Evidence from Tanzania." Journal ofEntrepreneurship, 2(1), 31-59.

Kessy, S.S. and F.M. Urio, 2006, The Contribution of MicrofinanceInstitutions to Poverty Reduction in Tanzania, Dar es salaam. Availableat: http://www.academia.edu/6391328/The_Contribution_of_Microfinance_Institutions_to_Poverty_Reduction_in_Tanzania.

Kipesha, E.F., 2013, "Performance of Microfinance Institutionsin Tanzania: Integrating Financial and Non financial Metrics."European Journal of Business and Management, 5(4), 94-105.

Kisaka, S.E., 2014, "The Impact of Attitudes towards Saving,Borrowing and Investment on the Capital Accumulation Process in Kenya:An Application of the Theory of Planned Behavior." Research Journalof Finance and Accounting, 5(9), 140-152.

Kline, R.B., 2011, Principles and Practice of Structural EquationModeling 3rd ed. D. T. Little, ed., New York: The Guilford Press.

Knowlden, A.P. and M. Sharma, 2012, "Health Belief ModelPredictors Of Undergraduate Students' Engagement In Short MessagingServices While Operating A Motor Vehicle." Safety Science Monitor,16(1), 1-10.

Krueger Jr., N.F., M.D. Reilly and A.L. Carsrud, 2000,"Competing Models of Entrepreneurial Intentions." Journal ofBusiness Venturing, 15(5), 411-432.

Lee, M., 2009, "Predicting and Explaining the Adoption ofOnline Trading: An Empirical Study in Taiwan." Decision SupportSystems, 47(2), 133-142.

Leung, K.-Y., C. Lo, H. Sun, and K. Wong, 2012, "FactorsInfluencing Engineering Students' intention to Participate inOn-Campus Entrepreneurial Activities." Journal of EntrepreneurshipEducation, 15, 1-20.

Lin, T.-C., J.S.C. Hsu and H.C. Chen, 2013, "CustomerWillingness to Pay for Online Music: The Role of Free Mentality."Journal of Electronic Commerce Research, 104(4), 315-333.

Mahlanza, T.J., 2015, Factors influencing retirement savingsintentions in Botswana. Deakin Univeristy. Available at:http://dro.deakin.edu.au/eserv/DU:30080443/mahlanza-factorsinfluencing-2015A.pdf.

Malebana, M.J., 2015, Perceived Barriers Influencing the Formationof Entrepreneurial Intention." Journal of Contemporary Management,12(1), 881-905.

Ministry of Agriculture Food Security and Cooperatives, 2015,Statement by Minister of Agriculture Food Security and Cooperatives Hon.Stephen M. Wasira (MP) During the IFAD Roundtable Workshop-TanzaniaCountry Programme Evaluation (CPE), Dar es salaam: United Republic ofTanzania. Available at:www.ifad.org/documents/10180/45751371-af2d-4196-96d4-0f8753b044b6.

Mishra, S., 2014, "Adoption of M-commerce in India: ApplyingTheory of Planned Behaviour Model." Journal of Internet Banking andCommerce, 19(1), 1-17.

Mohamed, A. and D. Ahmed, 2015, "The Challenges FacingMicrofinance Institutions in Poverty Eradication: A Case Study inMogadishu." International Journal of Humanities Social Sciences andEducation (IJHSSE), 2(2), 56-62.

Morduch, J. and B. Haley, 2002, "Analysis of the Effects ofMicrofinance on Poverty Reduction Analysis of the Effects ofMicrofinance on Poverty Reduction." NYU Wagner - Working PapersSeries, 1-163.

Morowatisharifabad, M.A., M. Momayyezi and M. Ghaneian, 2012,"Health Belief Model and Reasoned Action Theory in Predicting WaterSaving Behaviors in Yazd, Iran." Health Promotion Perspectives,2(2), 136-144.

Mukama, J., T. Fish and J. Volshenk, 2005, "Problems Affectingthe Growth of Microfinance Institutions in Tanzania." The AfricanFinance Journal, 7(2), 42-63.

Mwatsika, C., 2015, "Entrepreneurship Development andEntrepreneurial Orientation in Rural Areas in Malawi." AfricanJournal of Business Management, 9(9), 425-436.

National Bureau of Statistics, 2014, Basic Demographic andSocio-Economic Profile, Dar es salaam. Available at:http://www.tanzania.go.tz/egov_uploads/documents/Descriptive_tables_Tanzania_Mainland_sw.pdf.

National Bureau of Statistics, 2013, The United Republic ofTanzania Population Distribution by Age and Sex, Dar es salaam.Available at: ihi.eprints.org/2169/1/Age_Sex_Distribution.pdf.

Ng, B.Y., A. Kankanhalli and Y. Xu, 2009, "StudyingUsers' Computer Security Behavior: A Health BeliefPerspective." Decision Support Systems, 46(4), 815-825.

Phan, K.C. and J. Zhou, 2014, "Factors Influencing IndividualInvestors' Behavior: An Empirical Study of the Vietnamese StockMarket." American Journal of Business and Management, 3(2), 77-94.

Prabandari, S.P. and P.I. Sholihah, 2014, "The Influence ofTheory of Planned Behavior and Entrepreneurship Education towardsEntrepreneurial Intention." Journal of Economics, Business andAccountancy Ventura, 17(3), 385-392.

Preacher, K.J. and A.F. Hayes, 2008, "Asymptotic andResampling Strategies for Assessing and Comparing Indirect Effects inMultiple Mediator Models." Behavior Research Methods, 40(3),879-891.

Presser, S., M. Couper, J. Lessler, E. Martin, J. Rothgeb, M. Jean,and E. Singer, 2004, "Methods for Testing and Evaluating SurveyQuestions." 68(1), 109-130.

Putit, L. and Z.J. Johan, 2015, "Consumers "Acceptance ofHalal'' Credit Card Services: An Empirical Analysis."Journal of Emerging Economies and Islamic Research, 3(1), 1-9.

Rezaie, H. and D. Abadi, 2012, "Investigate theCustomers' Behavioral Intention to Use Mobile Banking Based on TPB,TAM and Perceived Risk (A Case Study in Meli Bank)." InternationalJournal of Academic Research in Business and Social Sciences, 2(10),312-322.

Rugumamu, C., 2014, "Empowering Smallholder Rice Farmers inTanzania to Increase Productivity for Promoting Food Security in Easternand Southern Africa." Agriculture and Food Security, 3(1), 7-15.

Safeena, R., H. Date, N. Hundewale, and A. Kammani, 2013,"Combination of TAM and TPB in Internet Banking Adoption."International Journal of Computer Theory and Engineering, 5(1), 146-150.

Sayid, O. and A. Echchabi, 2013, "Attitude of Somali Customerstowards Mobile Banking Services: The Case of Zaad and SahalServices." II(3), 9-16.

Schleider, J., A. Patel, L. Krumholz, B. Chorpita, and J. Weisz,2015, "Relation between Parent Symptomatology and Youth Problems:Multiple Mediation through Family Income and Parent-Youth Stress."Child psychiatry and human development, 46(1), 1-9.

Schlomer, G.L., S. Bauman and N. Card, 2010, "Best Practicesfor Missing Data Management in Counseling Psychology." Journal ofcounseling psychology, 57(1), 1-10.

Schumacker, R.E. and R.G. Loamx, 2010, A Beginners Guide toStructural Equation Modeling 3rd ed., New York: Taylor and FrancisGroup.

Shah Alam, S. and N. Mohamed Sayuti, 2011, "Applying theTheory of Planned Behavior (TPB) in Halal Food Purchasing."International Journal of Commerce and Management, 21(1), 8-20.

Shanmugam, A., M. Savarimuthu and T. Wen, 2014, "FactorsAffecting Malaysian Behavioral Intention to Use Mobile Banking WithMediating Effects of Attitude." Academic Research International,5(2), 236-253.

Shariff;, A.H.M., A. Amran and Y.N. Goh, 2012, "Determinantsof Behavioural Intention on Sustainable Food Consumption AmongConsumers' of Low Income Group: Empirical Evidence FromMalaysia." Journal of WEI Business and Economics, 1(1), 29-38.

Shinnar, R.S., O. Giacomin and F. Janssen, 2012,"Entrepreneurial Perceptions and Intentions: The Role of Gender andCulture." Entrepreneurship Theory and Practice, 36(3), 465-493.

Sobel, M.E., 1982, "Asymptotic Confidence Intervals forIndirect Effects in Structural Equation Models." Sociologicalmethodology, 13, 290-312.

Tanadi, T., B. Samadi and B. Gharleghi, 2015, "The Impact ofPerceived Risks and Perceived Benefits to Improve an Online Intentionamong Generation-Y in Malaysia." Asian Social Science, 11(26),226-238.

The United Republic of Tanzania President's Office PlanningCommission, 2009. A study on tranforming agriculture in Tanzania,Available at: www.mipango.go.tz.

Tingchi Liu, M., R. Chu, I. Wong, Z. Angel, Y. Meng, and C. Pang,2012, "Exploring the Relationship among Affective Loyalty,Perceived Benefits, Attitude, and Intention to Use Co-brandedProducts." Asia Pacific Journal of Marketing and Logistics, 24(4),561-582.,

Triodos Facet, 2007, Tanzania Country Scan Microfinance, AB Zeist.Available at: www.microfinancegateway.org/sites/default/files/mfg-en-paper-tanzania-country-scan-microfinance-2009.pdf.

Tsordia, C. and D. Papadimitriou, 2015, "The Role of Theory ofPlanned Behavior on Entrepreneurial Intention of Greek BusinessStudents." International Journal of Synergy and Research, 4(1),23-37.

Wani, S.N. and S. Malik, 2013, "A Comparative Study of OnlineShopping Behaviour: Effects of Perceived Risks and Benefits."International Journal of Marketing and Business Communication, 2(4),41-55.

Wolter, D., 2008, Tanzania - Why a Potential Food Exporter is stillImporting Food, Available at: www.oecd.org/dev/41302291.pdf.

Yuge, Y., 2011, The Current Situation of Microfinance inBangladesh: A Growing Concern about Overlapping Loan Problems-From aField Visit to Rajshahi and Comilla, Boston.

Julius Joseph Macha (a), Yee-Lee Chong (b), I-Chi Chen (c)

(a) Universiti Tunku Abdul Rahman, Faculty of Business and Finance,Malaysia [emailprotected]

(b) Universiti Tunku Abdul Rahman, Faculty of Business and Finance,Malaysia [emailprotected]

(c) Universiti Tunku Abdul Rahman, Faculty of Business and Finance,Malaysia [emailprotected]

Table 1 Fitness indices of the measurement modelParticulars GFI SRMR RMSEA CFIRecommended criteria >0.90 <0.08 <0.08 >0.90Initial model-index values 0.898 0.048 0.056 0.899Final model -index values 0.914 0.042 0.059 0.924Particulars NNFI Chi-square/dfRecommended criteria >0.90 <3Initial model-index values 0.885 2.538Final model -index values 0.908 2.705Note: GFI: Goodness of fit index; SRMR: Standardized root mean squareresidual; CFI: comparative fit index RMSEA: Root mean square of errorapproximation; NNFI: Non-normed fit indexTable 2 Construct validity and reliabilityParticulars PB Att SN PBC PBE BIAVE 0.50 0.59 0.51 0.51 0.56 0.58CR 0.80 0.88 0.80 0.76 0.79 0.73Note: Att: Attitude; SN: Subjective norms; PBE: Perceived benefits;BI: Behavioural intention; PB: Perceived barriers; PBC: perceivedbehaviour controlTable 3 Discriminant validityConstruct PB Att SNPerceived barriers (PB) 0.710 (*)Attitude (Att) -0.007 0.770 (*)Subjective norms (SN) 0.126 0.029 0.711 (*)Perceived behaviour control (PBC) 0.178 0.185 0.286Perceived benefits (PBE) 0.185 0.226 0.335Behavioural intention (BI) 0.040 0.278 0.465Construct PBC PBE BIPerceived barriers (PB)Attitude (Att)Subjective norms (SN)Perceived behaviour control (PBC) 0.713 (*)Perceived benefits (PBE) 0.414 0.747 (*)Behavioural intention (BI) 0.500 0.523 0.759 (*)Note: (*) represents the square root of AVE for each variable.Table 4 Path coefficients and hypothesis testing: Direct relationshipH IV Path DV Estimate (Std) S.E.H1 PB [right arrow] BI -0.101 (*) 0.029H2 PB [right arrow] Att -0.05 0.042H3 PBE [right arrow] BI 0.291 (***) 0.052H4 PBE [right arrow] Att 0.240 (***) 0.062H5 Att [right arrow] BI 0.154 (**) 0.037H6 SN [right arrow] BI 0.296 (***) 0.037H7 PBC [right arrow] BI 0.285 (***) 0.054H C.R RemarksH1 -1.964 SupportedH2 -0.934 Not SupportedH3 4.467 SupportedH4 4.286 SupportedH5 3.066 SupportedH6 5.048 SupportedH7 4.455 SupportedNote: Level of Significance: (***) p<0.001, (**) p<0.01, and (*)p<0.05. H: Hypothesis; IV: independent variable; DV: dependentvariable; C.R: Critical ratio; S.E: standard error; Att: Attitude; SN:Subjective norms; PBE: Perceived benefits; BI: Behavioural intention;PB: Perceived barriers; PBC: perceived behaviour controlTable 5 Path coefficients and hypothesis testing: Indirect relationshipHypothesis Path Estimate C.RH8a PB [right arrow] Att [right arrow] BI -0.008 -0.889H8b PBE [right arrow] Att [right arrow] BI 0.037 2.056Hypothesis BC-LB BC-UB RemarkH8a -0.029 0.008 Not supportedH8b 0.012 0.089 SupportedNote: BC: bias correction; LB: lower boundary; UB: upper boundary;C.R: Critical ratio; PB: Perceived barriers; BI: Behaviouralintention; Att: Attitude