The Scale Assessment and Validation of Microfinance Services and Household Socioeconomic Model: Using Parallel and Exploratory Factor Analyses

This study has followed a comprehensive methodology to develop and validate instrument for microfinance and household socioeconomic model. A complete stepwise explanation for questionnaire development is discussed based on methodology presented in Figure 1 below.

1.png

Figure 1. Methods diagram

2.1 First stage: Questionnaire development

The first stage of questionnaire development is referred to identification of theoretical support and essential factors to establish a conceptual model of this study. The ultimate purpose was to explore all related variables and develop their measuring items. Hence, an extensive literature review and discussion with team members were conducted to identify related theoretical foundation and crucial factors improving households’ socioeconomic performance.

2.1.1 Theoretical foundation

The Household Economic Portfolio (HEP) Model developed by Chen and Dunn [26] was found to be the most widely accepted theoretical work in the literature of microfinance impact assessment. Hence, this study has used the theoretical model of HEP as foundation to develop its framework. The HEP model discourses the question that how micro-financial interventions impact various households’ outcomes. According to the HEP model, a household’s economic mechanism consists of economic resources, set of economic activities, and interconnection between them. The HEP model theorizes that financial interventions directly contribute to the set of households’ economic resources which further influence socioeconomic outcomes through set of certain economic activities.

The HEP model proposes that financial and social are the two major external factors that contribute to the households’ resources and indirectly influence their socioeconomic outcomes through certain actions such as investment, production, and consumption activities. In the HEP framework, financial capital is considered as major input factor that can help households to enhance their socioeconomic performance [20]. Mainly, allocation of financial capital to different economic activities is influenced by various factors such as decision-making structure, preferences, opportunities, and constraints [27]. Likewise, allocation of financial capital has significant consequences for the effectiveness of financial interventions in achieving improved socioeconomic outcomes. For example, allocating funds to investing and production activities can boost economic performance, whereas spending more on consumption activities does not provide similar outcomes.

In this regard, the HEP model considers allocation of financial capital towards different economic activities as a crucial phenomenon in improving socioeconomic performance. This indicates the importance of decision-making process that is determined by several individual and household specific attributes in enhancing socioeconomic outcomes. Likewise, the HEP model also highlights the importance of external market factors in improving the success of microfinance interventions. However, the HEP model has a limitation as it does not provide sufficient detail on such factors that might shape households’ decision-making and influence their socioeconomic outcomes. In this regard, the current study has introduced human capital factors in the HEP model to explore their role in enhancing the effectiveness of microfinance interventions. In addition, this study also contributes to the HEP model by introducing microfinance institutions’ service efficiency as institutional factor in the household economic framework.

2.1.2 Developing theoretical and conceptual framework

Following the theoretical literature, the theoretical (initial) structure of the framework was obtained which consisted of (1) microfinancing as independent variable with two dimensions that were financial literacy (financial services, financial management, and social capital) and entrepreneurship (entrepreneurial competencies and entrepreneurial environment), (2) Institutional factors as moderation with two variables of microfinance’ efficiency and government support policy, and (3) household socioeconomic well-being as dependent variable with two dimensions that are social well-being and economic well-being (refer Figure 2).

2.png

Figure 2. Theoretical framework (Initial)

Next, an extensive discussion was conducted with industry players to gain more details and confirmation on research areas (variables). Hence, revisions were made to develop the final conceptual framework. Mainly, two new independent variables were included, such as TP and BC. Similarly, FMP and EC were reclassified as mediation variables whereas variables of social capital, entrepreneurial environment, and government support policy were dropped from the model. At the same time, ES was also added to the existing dependent variables. Thus, new research areas of Quality Educational Provision framework presented in Figure 3 were completely endorsed by the experts. The final developed model includes three main microfinance services such as FS, TP, and BC as independent variables. Next, two human capital factors which are FMP and EC are incorporated as mediation variables to explore the impact pathway of microfinancing towards improved socioeconomic outcomes. Then, considering the importance of service delivery, MIE is considered as moderation variable that might play vital role in enhancing the effectiveness of microfinance. At last, household socio-economic performance is included as dependent variable with three dimensions, that are SW, EW, and ES (refer Figure 3). Overall, this study implies that microfinance financial and non-financial services help clients to develop their capabilities such as entrepreneurial and financial management skills. These capabilities ultimately assist them to make effective decisions in selecting and managing their economic activities, thus, resulting in improved socioeconomic outcomes. The list of these variables and their respective dimensions are presented in Table 1.

3.png

Figure 3. Conceptual framework of quality educational provision

2.1.3 Items development

This study developed questionnaire items in four steps. First, measuring items were acquired from the literature for all identified variables. Then, industry experts were consulted to ensure relevancy of the items (statements) and specific dimensions of the framework. Next, statistical methods were applied to examine validity of the constructs. At last, all developed items were cross-checked against their concepts and objectives of this study.

In this regard, a questionnaire used as instrument was initially designed by acquiring all the items (statements) for each dimension from the initial framework. The process of adoption (i.e. retain original items) and adaption (i.e. change the actual items) approaches from literature were used to create all the items. At the initial stage, this study developed a total of two hundred six items from various existing studies and authors’ discussions. In terms of the principle of wording, several principles had been considered such as no double barrel, no ambiguous statement, no bias (race or group), and appropriate length of the questionnaire (ideal answering time). Likewise, the criteria of minimum three items for each dimension was also followed in developing items.

Then, after extensive discussion with industry representatives, this study developed final framework and concepts presented in the Figure 3. Consequently, 110 items were deleted from 206 of the original items, and then 30 new items were added; hence, leaving a total of 126 items in the questionnaire. Then, various statistical tests were employed to assess quality and authenticity of the measurements. These data analysis steps are discussed in the next section. After conducting statistical methodology, final 125 items are validated in this study. The detail of these items is discussed below whereas actual items are listed in the Table 4 of the results section.

The first independent variable of the study i.e. FS was measured using three main financial services of five items each. These services include microcredit, micro-savings, and death-benefit fund. The measuring items were adapted from the existing literature [22]. Then, second independent variable i.e. TP was measured on the basis of basic training as its major dimension. Total 5 items were developed and designed following the industry inputs and program features. Similarly, third independent variable i.e. BC was captured on the basis of two classifications; service coaching and product development coaching. It was measured using a total of 10 items developed from the industry input.

Next, first mediating variable of FMP was measured using five major financial management practices found in the literature including basic finance, budgeting and cash management, credit management, risk management, and capital accumulation. These practices were measured by adapting 27 items from the various existing studies [28-32]. Likewise, next mediating variable of the model i.e. EC was measured using six dimensions given by Man et al. [33]. These include commitment competency, conceptual competency, organizing competency, relationship competency, opportunity recognition competency, and strategic competency. Total 33 items were adapted following the previous literature [33] to measure EC.

Table 1. Variable description

Then, this study measured its moderating variable i.e. MIE, using two main dimensions of credibility, and responsiveness. Total 14 items were developed by adapting from the study of [25]. Further, dependent variable of the study that is household socioeconomic performance was classified into three major dimensions including SW, EW, and ES. Specifically, EW was measured by developing 5 items of different household economic performance outcomes such as income level, asset possession, expenditures, savings, and confidence. The SW was measured using two social development dimensions; level of satisfaction and provision of opportunities. Following the existing literature [34], 7 items were adapted to measure satisfaction level in meeting basic needs, and 3 items were used to measure provision of opportunities. Finally, ES was measured using 7 items of business performance developed and adapted from the previous literature and industry input [22].

2.2 Second stage: Questionnaire testing (data analysis)

In the second stage, this research conducted a detailed data analysis to test questionnaire’s reliability and validity. First, reliability of the data was assessed for all the measuring constructs. At the same time, validity of the constructs was also examined. For this, factor analysis and parallel analysis techniques were applied to extract the significant number of factors and explore underlying patterns of the variables. Lastly, reliability of the developed scale was again assessed in this stage. A detail of various steps conducted at this stage is given below.

2.2.1 Content validity

The initial process in the second stage began with content validation assessment when a group of experts from industry and academic was sought to assess the instruments. Based on the assessment, the content validity index (CVI index) has been calculated using a 4-point ordinal rating scale ranging from irrelevant to very relevant (1= Irrelevant, 2 = Not Important, 3 = Relevant with correction, and 4 = Very relevant). The scales of one and two represent disagreement, while scales of three and four represent agreement. All 126 items received a score above 96%, indicating congruence between the six reviewers. Hence, no items were deleted, while items revision was conducted to alleviate inaccurately and out-of-context sentences.

2.2.2 Pretesting

Under the pretesting step, the subsequent version of the questionnaire was pretested on six respondents of AIM clients (sahabat) from different districts of Negeri Terengganu Malaysia. From the respondents' feedback, the questionnaires are reorganized as follows: Part A consists of Demographic information, and Part B consists of structured and semi-structured questionnaires. Since no item being deleted in the pretesting step, 126 items were maintained. In this step, any comments and suggestions were obtained to ensure the clarity, relevancy, an understanding of the statement, vague sentences, and answering time of the questionnaire.

2.2.3 Construct reliability and validity (pilot study)

Examining construct validity is a crucial step in the process of scale development. It refers to the extent to which measuring constructs represent their respective theoretical concepts [35]. Hence, this study conducted factor and parallel analyses techniques to examine construct validity. A survey method was used as it is considered more appropriate to collect large amounts of quantitative data.

Respondents and sample size: The target population is comprised of all B40 households receiving microfinance services offered by AIM in Malaysia. Regarding the sample, AIM clients (sahabat) were chosen as respondents from all Malaysian states to ensure that results represent the population. The convenience sampling approach was adopted to collect the data. Determining sample size is one of the important steps while conducting an empirical study. Usually, the sample size is determined based on the minimum sample required to achieve reliable results. Particularly, the problem of sample size is more crucial in case of factor analysis because it involves greater level of subjectivity [36].

This study determined its sample size following the stable factor structure (SFS) recommendations. The SFS recommends sample size based on two crucial aspects that are mathematical overdetermination and size of communalities. In order to reduce standard errors of correlations, SFS suggests a sample size of at least 100 to 200 and respondent variable ratio of 2 to 1 [37, 38]. Hence, a total of 200 respondents were selected using a convenient sampling method. Data were collected from respondents through telephone calls because of COVID-19 related restrictions. All data were collected anonymously, and all confidentiality and privacy were maintained. In addition, this study also conducted a manual filtering and screening to remove any inappropriate questionnaire scripts from the collected data. The next step is performing the data analysis.

Reliability analysis: Before performing factor analysis, this study examined the reliability of the data. For this, four commonly used criteria are followed to ensure the reliability of the data. These criteria include testing standard deviation (SD), inter-item correlations, corrected items total correlations, and internal consistency using Cronbach alpha. Hence, the reliability of data is ensured as per the following conditions:

1. Standard deviations are greater than zero;
2. Inter item correlations are between 0.3 to 0.9;
3. Corrected items total correlation is greater than 0.3;
4. Cronbach alpha is greater than 0.7.

Factor analysis: This research employed the factor analysis method, an important statistical technique wisely adopted for scale development. This method has three primary uses. First, it is employed to reduce a large set of variables into a small number of factors. Next, it facilitates the formation of the model by determining underlying associations between latent variables and their measurements. At last, it also examines the validity of the scale. As the purpose of this study is to develop scale and model; therefore, it applies an exploratory method of factor analysis.

This research has conducted exploratory factor analysis (EFA) for developing a questionnaire to explore underlying patterns of variables. This EFA test also was conducted to reduce larger data set into smaller sets of variables and to discover underlying theoretical structures. Three comprehensive steps involved: factor suitability, factor extraction, and rotation.

First step, factor suitability test was conducted. Existing literatures indicate that most researchers ignored assessing factor suitability before conducting factor analysis [39]. However, testing factor suitability is crucial as it confirms that factor analysis should be applied to the data. Therefore, this study examined factor suitability by conducting Kaiser-Meyer-Olkin (KMO) test [40] and Bartlett’s test [41]. The threshold values for these tests are as follow; KMO value should be greater than or equal to 0.6, and Bartlett test’s probability value should be significant at 0.05 or less [2].

Second step, factor extraction was conducted. For this purpose, Principle Component Analysis (PCA) was conducted, which is the most widely adopted method. Most existing studies suggest using multiple criteria for determining and retaining a total number of factors [42]. Most commonly used methods include Kaiser’s eigenvalue criteria [43] and Cattell’s scree test [44]. It is observed that both methods are being widely adopted in the research because of their simplicity and easiness. However, both methods have been criticized because of certain serious deficiencies in their performance [42]. On the other hand, Horn’s Parallel Analysis (PA) [45] has been developed to overcome these deficiencies. PA involves a complex procedure; therefore, it has less popularity among researchers. However, PA is considered the most appropriate method for determining total number of factors [42]. In this study, a modification of Horn's parallel analysis based on Monte Carlo simulation is used to improve the analysis further and accuracy of the results.

Hence, for factor extraction, this study has applied all three methods, i.e. Kaiser’s eigenvalue criteria, Cattell’s scree test, and Horn’s parallel analysis, to determine a total number of factors to be retained. For Kaiser’ Criterion, best fitted line was determined on the basis of largest sum squared of distance (SSD). Then, Singular Value Decomposition (SVD) method was used to scale the line into one unit. Hence, this transformation (new results with same proportion) is termed loading proportion (eigenvector). These eigenvectors can be viewed as weights of the variables. The highest eigenvalue related with each eigenvector shows substantive importance for each variable or component. Under Kaiser Criterion, the basic rule is to retain factors with eigenvalue of greater than 1 and ignore factors with eigenvalue less than 1 [43].

Next, this study also incorporated Scree test to determine and retain total number of factors. The Scree test proposed by Cattell is a well-known factor retention method which refers to the visual examination of eigenvalues. In this method, eigenvalues are graphically represented in the descending order where breakpoint is used to determine retaining factors. The rationale behind this test is that breakpoint on the graph separates the trivial factors from non-trivial factors [44]. It indicates that the steep part of the graph before the breakpoint shows major factors which account for the most of variance whereas after breakpoint represent those minor factors which relatively account for smaller variance.

Under Scree Plot test, this study reconfirmed the results from Kaiser’s criterion by plotting eigenvalues on y-axis and their respective components on x-axis. For illustration purpose, an example of Scree plot is shown in Figure 4 [44]. Hence, the number of components and variables are determined based on points that appear to be on the left side of inflection (also called as elbow); however, the inflection point itself is not to be considered. In the given illustration, two points i.e. 1 and 2 can be considered for selection.

While Horn’s Parallel Analysis (PA) [45] was performed on all items to double check results from Kaiser’s criterion and scree plot by producing more accurate and reliable factors. Literature suggests that Horn’s parallel analysis is a sample-based method that provides superior results to other widely used techniques such as eigenvalue criteria and scree test [42]. This study has developed a simple approach, adapting from the previous literature to conduct PA using SPSS [46]. PA is undertaken in four simple steps shown in Figure 5 below. In the first step, this study has generated random data of the same number of variables and sample size as the actual observed data. And this is to comply with the condition that a random dataset must have a similar number of variables and observations. In the next step, principle component analysis was performed on the randomly generated data to compute eigenvalues.

4.png

Figure 4. Scree plot (developed based on Field, 2013) [38]

5.png

Figure 5. Steps to conduct parallel analysis (Monte Carlo Simulation)

Further, Horn [45] suggests that the number of random data matrices must be of sufficient size to avoid biasness. Hence, following the existing literature, both the first step and second step were repeated fifty times to collect adequate observations. Then, in the next step, mean eigenvalues and 95th percentile eigenvalues of fifty random data sets were computed with average eigenvalue and 95th percentile value for each factor. In the last step, eigenvalues of actual observed data were compared with these computed mean eigenvalues and 95th percentile eigenvalue of randomly generated data. For instance, the first real eigenvalue is compared with the first mean eigenvalue from random data, and then the second real eigenvalue is compared with the second mean eigenvalue. However, this comparison could be easily performed by comparing scree plots. Under this method, a factor is selected if its eigenvalue of actual data is greater than the mean eigenvalue and 95th percentile eigenvalue of randomly generated parallel data. Moreover, the number of factors to be considered also relies on the basic objective of the analysis.

Third step is factor rotation. The Direct Oblimin method is incorporated for producing correlated factors. This study retained factors/items on the basis of factor loadings from this process. Hence, authors retained factor/items with factor loading of 0.3 and above whereas redundant items between the components (cross loading) were deleted. Under Direct Oblimin method, two tables have been conferred, i.e. Pattern Matrix and Component Matrix (Factor Transformation Matrix Table and Varimax/Orthogonal – Rotated Factor Matrix Table). A detailed assessment of both tables was performed to select components and items.

At this stage, the following procedures were used to do factor rotation and to complete the EFA process in this study:

1. Choose the more interpretable table – good grouping and most items (pattern and structure matrix table);
2. Delete redundant (cross-loading) items between components;
3. Delete items with factor loading lower than 0.3;
4. The item needs to be deleted one by one and re-run the EFA test;
5. If the result improves, no further deletion is needed, and
6. Finally, Parallel Analysis (PA) will be conducted to conform results from Kaiser.

After the required validity is achieved for all the factors and components, this study conducted final test, i.e. to determine reliability of the developed scale. Reliability testing was performed to examine the consistency between items/attributes [47]. The purpose behind this test is to ensure reliability of indicators for measuring their respective latent variables. Specifically, Cronbach’s Alpha values were used to assess reliability of the factors.

2.3 Third stage: Questionnaire cross checking

The last stage in the process of instrument development refers to the verification of the results. Particularly, final factors are cross-examined at this stage to confirm their accuracy and efficiency. Hence, after proving validity and reliability of the constructs through statistical analysis, this study also conducted questionnaire cross checking in the final stage to ensure further accuracy for the final data collection. Questionnaire items cross checking is a crucial step to avoid mistakes and achieve accuracy in the results. In this regard, all final retained items of the instrument were cross checked against their specific theoretical concepts and research areas identified by the experts. The purpose is to prove that the variables sufficiently reflect their conceptual meanings. It would indicate that all the developed constructs and their items incorporate their respective research areas of the specific variables. This study identified four main research areas including; 1) microfinance effectiveness, 2) human capital development, 3) service efficiency, and 4) sustainable socioeconomic development. Remaining 125 items covering eight measuring constructs were cross-examined by the researchers against these specific research areas.

For instance, FS was cross-checked against facilitating clients by providing specific financial benefits. The strength of non-financial services such as TP and BC is determined on the basis of ability to enhance human capabilities of their clients. All together both financial and non-financial services mainly address effectiveness of microfinance interventions. Whereas, non-financial services also play a vital role in addressing human capital development. Likewise, the mediating variables such as FMP and EC have strength in providing benefits to the micro-entrepreneurs in form of their improved capabilities. As adoption of FMP and improved EC can help clients to efficiently manage their economic matters. Thus, these both constructs measure human capabilities which further contribute in enhancing households’ socioeconomic outcomes. Furthermore, the items of MIE were cross-examined with their ability to capture the fact that how efficiently microfinance institutions are providing their services to the clients. Hence, it specifically addresses efficiency of microfinance service providers and play vital role in the effectiveness microfinance to achieve sustainable socioeconomic development. At last, items of households’ socioeconomic performance indicators (SW, EW, and ES) address a broader research area of sustainable socioeconomic development covering households’ social status and economic performance. In addition, all steps of instrument development and validation were also revisited to confirm accuracy of the results. Hence, after this stage, final validated and reliable instrument for the specific variables are reported in this study.

The primary purpose of this study is to develop a new measuring tool for microfinance and household economic portfolio model in Malaysia. Therefore, the research aims to develop and validate the factor structure and dimensionality of the constructs in microfinance and household economic portfolio model. This study has followed standard methodology that involves an extensive review of existing constructs and appropriate statistical analysis. Specifically, EFA is employed to validate and assess the instrument's factor structure.

Considering the longstanding issue of humankind’s economic deprivation and poverty, authorities continue to search for effective tools that can help in uplifting the socioeconomic status of poor individuals. Similarly, researchers continue to investigate various socioeconomic interventions and their role in enhancing individuals’ socioeconomic performance. Hence, this research has significant contribution towards a newly developed microfinance and households’ socioeconomic model, which emphasizes assessing the role of various microfinance services and household specific factors in eradicating poverty. In addition to developing a reliable measuring instrument for this microfinance and household socioeconomic model, this research has uncovered a number of crucial variables that could significantly help to improve households’ socioeconomic status and reduce poverty in the world. Based on the EFA’s result, the study found that there are seven factors namely financial services (FS), training programs (TP), business coaching (BC), financial management practices (FMP), entrepreneurial competencies (EC), microfinance institutions’ service efficiency (MIE), and households’ socioeconomic performance (SW, EW, and ES) that meet the reliability and validity test.

Moreover, to ensure further accuracy, all final items covering eight constructs were cross-checked against research areas suggested by the industry experts. Microfinance factors were cross-checked against their effectiveness in helping clients to avail maximum benefits. For instance, the final factor structure indicates that all items under microcredit and micro-savings represent financial services’ accessibility and adequacy, which helps poor clients promptly meet their financial needs. Similarly, items under the death-benefit fund cover contribution aspects and claim delivery, influencing the client’s participation in the program. Next, all items under constructs of basic entrepreneurship, service coaching, and product development; capture the effectiveness of training and coaching programs in developing capabilities among clients to improve their business performance. 

Furthermore, the variables of financial management practices and entrepreneurial competencies were cross-checked against the research area of human capital development. Here, the items under various constructs of financial management measure the extent to which clients were able to manage their financial matters after participating in microfinance programs. Likewise, items under the constructs of entrepreneurial competencies measure improvement in clients’ capabilities relating to enterprise management and entrepreneurial strategies. In addition, the constructs of microfinance service efficiency have significant role in the success of microfinance services to help uplift poor clients. In this regard, the items under the construct of credibility represent the client’s level of trust in the microfinance institutions and their employees. Similarly, the items under the construct of responsiveness represent the response efficiency of microfinance institutions toward clients’ problems.

Finally, the constructs of households’ social wellbeing, economic wellbeing, and entrepreneurial success were cross-checked against the concept of sustainable socioeconomic development. Expressly, the items under social wellbeing represent clients’ level of satisfaction in meeting their social needs after participating in microfinance schemes. In a similar manner, the items under economic wellbeing and entrepreneurial success represent improvement levels in clients’ economic and business performances, respectively.  

Consequently, given all discussion regarding final items and their respective constructs, this study shows that the developed instrument is consistent with the industry experts' research areas. Hence, this instrument is proved valid and effective to be applied in future related research.  

4.1 Contributions to the literature

This research provides significant contributions to the existing body of literature. Most importantly, it offers an instrument to assess the impact of microfinance services on household socioeconomic performance. This instrument is proved valid and reliable in measuring various important microfinance and household specific factors. Hence, this study opens great opportunities for the researchers to incorporate these constructs in the future empirical research. Results from this study provide measuring constructs for three types of microfinance services including FS, TP, and BC. Existing literature in Malaysia [10, 19] has mostly focused on microfinance financial interventions and trainings. Hence, this research introduces a new explanatory factor that is business coaching (BC) in the microfinance and household economic model. Introducing business coaching service as input factor would have significant implications in the context of human capital development in Malaysia. Particularly, business coaching could be a vital factor for low-income business holders as such individuals usually have insufficient capabilities to compete in the market as mentioned in the past studies that business coaching is an effective tool in enhancing business performance [48]. Hereafter, provided with proper consultation and support from microfinance coaches, microbusiness holders could develop effective skills and strategies to achieve their business goals. 

Furthermore, both mediating factors; FMP and EC are also proved valid and reliable. Most of the existing empirical studies [4, 10] are limited to the microfinance impact assessment, thus, ignoring the underlying mechanisms that can explain the impact pathway towards improved socio-economic outcomes. Hence, this research highlights two potential mediating factors; entrepreneurial competencies and financial management practices explaining the impact pathway from microfinance to household socioeconomic performance. Entrepreneurial competencies have been discussed vastly in the entrepreneurship literature [49]. Same as financial management practices in the relationship to SME business performance [50]. Thus, these both factors have significant implications for the authorities and low-income households. This research identifies these both factors as empowering capabilities that could help in enhancing the effectiveness of microfinance to eradicate poverty.

In addition, this research also provides a valid and reliable measuring construct for microfinance institutions’ service efficiency. Introducing this new factor as potential moderator in the microfinance and household economic framework is another major contribution in the existing body of knowledge. Considering the dual objective of microfinance institutions, they have to be more efficient in delivering services for the betterment of their participants [25]. Therefore, service efficiency of microfinance institutions is expected to play crucial role in helping households to achieve their social and financial goals. However, the literature has overlooked the role of microfinance institutions’ service efficiency in the contact services effectiveness to enhance socio-economic performance [20]. As majority of the existing research has only focused on technical efficiency measurements for the microfinance institutions. Thus, leaving a gap regarding the measurement of service delivery efficiency and its role in the microfinance mechanism. It is implied that successful utilization of financial support strongly depends on efficient service delivery [25]. Therefore, prioritizing efforts toward service delivery will ultimately enhance microfinance clients’ wellbeing and economic performance. At last, considering multidimensional aspects of a household wellbeing, this research also provides three diverse measuring constructs including social wellbeing (SW), economic wellbeing (EW), and entrepreneurial success (ES). These different indicators would help in better understanding the socioeconomic effectiveness of microfinance.

Overall, this research is a significant contribution to the existing microfinance research which lacks in providing established constructs [22]. Most of the studies [3, 12] have used diverse objective measures such as length of participation in microfinance, amount of loan, and training hours as proxies for the microfinance services. Whereas, few studies [19, 23] have used subjective measures for the microfinance factors in the impact assessment. Hence, by providing a valid and reliable measuring instrument, this research opens multiple opportunities for the future researchers to explore microfinance sector of Malaysia and its role in eradicating poverty.

4.2 Theoretical and practical implications

In terms of theoretical implications, this study contributes to the HEP model by identifying the impact channel of microfinance through human capital development. It implicates that micro-financial intervention directly contributes to a household's financial and human capital, and mainly human capital development plays a crucial role in effective decision making, which ultimately leads poor households toward improved socioeconomic outcomes. Likewise, it also highlights microfinance institutions' service efficiency as an external market factor in the HEP model that could enhance the effectiveness of financial interventions in achieving social goals. Based on the factors from the model, this research implies that microfinance services assist households in developing effective human capabilities such as entrepreneurial competencies and financial management practices. In turn, these capabilities help households efficiently manage their financial and business operations, thus, enhancing socioeconomic performance. However, this theoretical mechanism highly depends on the service efficiency of microfinance institutions.

Further, this study also provides several practical implications for the authorities, microfinance institutions, and households. For instance, this study implicates that authorities should focus on developing human capabilities among poor households through effective training and educational strategies. In summary, authorities should emphasize on enhancing the financial literacy of poor communities, that would help them efficiently utilize their limited resources. Next, microfinance institutions are suggested to strengthen their non-financial services further as these services play a crucial role in developing managerial capabilities among their clients. Meanwhile, microfinance management should also develop strategies such as strong monitoring mechanisms to improve employees' efficiency in delivering services to needy households. This would help their clients in availing productive economic opportunities on time. Moreover, low-income households are also suggested to improve their business and management skills and financial awareness, which would ultimately help them achieve long-term socioeconomic goals.