1. Introduction
Poverty has become a multidimensional problem in developing and underdeveloped countries. According to the
World Bank (
2018), 2.4 billion people live in an extremely poor condition spending less than US
$1.90 each day. Specifically, poor low-income households are unable to invest in education and generate opportunities due to limited access to credit and financial support. Therefore, various organizations such as United Nations (UN), World Bank, and non-governmental organizations (NGOs) have been introducing different programs, services, and policies to address poverty issue (
Akoum 2008). Of all these initiatives, financial assistance is the utmost importance.
Nichter and Goldmark (
2009) indicated a strong relationship between micro-credit programs and income, which is likely to decrease economic vulnerability (
Al-Mamun et al. 2014).
Most poverty measurements are introduced based on the poverty line that distinguishes the income and expenditure between low-income, poor, and non-poor households (
Hagenaars and Praag 1985). Although income is an important measurement of poverty, it is not effective to capture the poverty (
Chatterjee et al. 2014). Poverty is a multidimensional phenomenon with a richer concept than the traditional approach (
Chatterjee et al. 2014). In Malaysia, poverty is measured by using poverty line income (PLI), a quantitative money metric measure. A household is considered poor when it’s monthly gross income is below the PLI. Basically, households with a gross monthly income below RM760 and RM460 in Peninsular Malaysia are absolute poor and extreme poor, respectively. The low-income households group is classified based on an average monthly household income of 40% in Malaysia, which is RM2848 (
DOSM 2017).
Poverty eradication is one of the main objectives that needs to be addressed by the development policy. Therefore, micro-credit and training programs are included in the New Economic Policy (NEP) to alleviate poverty and income inequality. To improve the socio-economic conditions of low-income households, several strategies and policies are proposed such as the NEP, National Development Policy, Vision 2020, and New Economic Model (NEM), reducing the incidence of poverty from 0.6% in 2014 to 0.4% in 2016 (
EPU 2018). Unfortunately, poverty has been prevalent that requires organizations such as Amanah Ikhtiar Malaysia (AIM), Tabung Ekonomi Kumpulan Usahawan National (TEKUN), and Lembaga Kemajuan Ikan Malaysia (LKIM) to lessen economic vulnerability and poverty. These three non-governmental organizations provided micro-credit and training programs for the low-income household to generate income.
Undeniably, micro-credit and training programs are prominent tools for poverty eradication in Malaysia. The government can collaborate with AIM, TEKUN, and LKIM by providing low-income households working capital and useful programs to generate income (
Ahmed et al. 2011;
Md Saad 2011;
Al-Mamun et al. 2011) and decrease the level of economic vulnerability (
Al-Mamun et al. 2014). Basically, AIM is a private trust established on 17 September 1987 registered under the Truster (Incorporration Act 1952). It provides loans such as I-Mesra Loan, I-Srikandi Loan, I-Wibawa Loan, and I-Penyayang Loan. AIM provides I-Bestari Loan for education and I-Sejahtera Loan for housing and other purposes. Moreover, AIM also provides non-financial assistance programs such as Sahabat Teras, Strengtening Entrepreneurship, and Sahabat Berjaya. These programs include the understanding of skills and risk management plans.
Apart from that, TEKUN is an agency under the Ministry of Entrepreneurial and Cooperative Development established in November 1998. TEKUN provides four economic loans including TEKUN Nasional Financial Scheme, Teman TEKUN Financing Scheme, I-Factoring Financing Scheme, and Ar-RAHNU TEKUN. LKIM is an authorized body under the Ministry of Agriculture and Agriculture Base Industry incorporated under Malaysia Fisheries Development Board Act 1971. LKIM provides several small-scale working capital services to the fishing community. LKIM offers one funding section for loan, which is to plan and coordinate the fisheries funds and economic development fund of fishing for fishermen. All in all, the main purpose of establishing these three organizations is to reduce the poverty issue among low-income households by providing various development activities that assist them in generating more income.
Development initiative programs are the backbone of many developing and emerging economies which hold the key to the possible revival of economic growth and the elimination of poverty (
Afrane 2002). In Ghana and South Africa,
Afrane (
2002) reported that a significant portion of the increased income was channeled into improving access to financial outcomes (savings and accumulation of assets) and non-financial outcomes (education, health, standard of living, housing, and job opportunity) (cf.
Hossain and Knight 2008;
Odell 2010). Also reported in South Africa,
Hietalahti and Linden (
2006), found that a great proportion of micro-credit participants were able to secure profit of more than
South African Rand (ZAR) 50 per month, where half of the participants secured profits of more than ZAR500 per month. In Zimbabwe,
Barnes and Keogh (
1999) showed that the participant’s household income more than doubled.
In investigating the impact of Malaysia’s development initiative programs,
Omar et al. (
2012) found that participants of micro-credit programs had seen an increase in average monthly household income, from RM1286.77 to RM2703.63, an increment of RM1416.86 (110%), more than double the previous amount. According to
Samer et al. (
2015), existing members of the programs enjoyed a higher household income, an increase of 1.5%, as compared to new members. This was because the existing members had spent over three years in the programs with frequent participation in business training.
However, few studies argued that the impact of development initiative programs, in particular, micro-credit and training programs, have been inconclusive (
Angelucci et al. 2013;
Ganle et al. 2015;
Van Rooyen et al. 2012). According to
Ganle et al. (
2015), the participants underscored the limitations of small-sized loans that put high profit-yielding business venture investments out of reach as they require larger capital investments. Consequently, some participants failed to invest their loan in economically rewarding ventures, are faced with considerable loan repayment problems, and had even become more vulnerable.
Angelucci et al. (
2013) further broaden the lens by examining five measures of additional income generated by micro-credit programs in the last six months—total household income, labor income, participation in any economic activity, remittance income, and positive savings—and revealed no significant effects on any of the five measures. Similarly, another study conducted in the Sub-Saharan African region showed micro-credit doing harm as well as good to the poor people it purports to help. Thus, the researchers concluded that micro-credit should not be promoted as the absolute solution due to the potential harm towards the poorest participants (
Van Rooyen et al. 2012). Generally, instead of focusing on the hardcore and extreme poor cohort, some micro-credit programs only focused on benefiting the cohort of poor (
Altay 2006;
Copestake et al. 2001;
Rahman and Razzaque 2000) by assisting in utilizing their money (
Rutherford 1996). In addition, the majority of these studies were conducted in one geographical area (
Omar et al. 2012;
Saad and Duasa 2011), within an organization (
Mahmood and Mohd Rosli 2013;
Al-Mamun et al. 2011;
Nawai and Bashir 2009), using small sample size (
Ganle et al. 2015;
Omar et al. 2012), and limited control variables (
Hashemi et al. 1996).
To reiterate, this study extends the literature by examining the impacts of access to working capital and training programs on households’ income and economic vulnerability among the participants of development initiatives in Kelantan, Malaysia. Rather than proving the positive path of micro-credit and training programs towards household income, we seek to ascertain whether the participants experience a decrease in economic vulnerability level after gaining access to the said programs. Kelantan was chosen as it is the poorest state in Peninsular Malaysia, with 0.4% poverty rate in 2017, and scored the lowest mean monthly household income of RM4214 (
EPU 2018).
3. Research Methodology
This study used a cross-sectional design and collected quantitative data through structured interviews. The population of this study is a total of 88,435 low-income households identified as participants of development programs offered by AIM, TEKUN, and LKIM in Kelantan, Malaysia. The research team approached the said development organizations for a list of at least 150 participants, each with their name, address and contact details. AIM, TEKUN, and LKIM provided a list of 500, 350, and 156 randomly selected existing participants of their programs. The listed participants (1006 participants) were from seven districts including Tumpat, Bachok, Pasir Puteh, Pasir Mas, Tanah Merah, Gua Musang, and Jeli. Then, the research team communicated with each of the listed participants to explain the purpose of the survey and arrange for an appointment with them. Of the 1006 listed participants, this study secured the participation of 450 respondents (AIM-150; TEKUN-150; LKIM-150). Data was collected from the respondents through structured face-to-face interviews conducted at their preferred location.
3.1. Sample Size
s = required sample size.
= the table value of chi-square for 1 degree of freedom at the desired confidence level (3.841).
N = the population size (88,435).
P = the population proportion (assumed to be 0.50).
= the degree of accuracy expressed as a proportion (0.05).
Since the population was 88,435, a sample size of 383 was required. To minimize possible complication regarding small sample size, this study collected data from 450 participants.
3.2. Operational Definitions
Length of participation refers to the duration respondents spend on the participation in micro-credit and enterprise development training programs. The total amount of economic loan received refers to the amount of credit that participant obtain from AIM, TEKUN and LKIM. Hours of enterprise development training refers to the total number of enterprise development training programs attended, total number of training hours attended, and total number of attendants in the last 12 months. Household income refers to ‘average monthly income obtained from all sources by all the household members in last twelve months’. Finally, economic vulnerability refers to the risk of exposure to potentially harmful events. Studies conceptualized vulnerability as vulnerability to income poverty, asset poverty, and the risk exposure to political, natural, and economic disasters. In fact, economic vulnerability is measured by using the index adopted from
Al-Mamun et al. (
2014), presented below.
refers to the vulnerability index that measures the level of economic vulnerability among the participating households.
denotes the coefficient of variation for the average monthly income earned (last twelve months) among the three groups of households based on their business period.
, where
represents the average net worth of enterprise assets within the same group of respondents, while
reflects the net worth of enterprise assets.
is the proportion of total income from enterprise income (businesses owned and managed by the respondents). Meanwhile, the effect of poverty level upon economic vulnerability was measured as
, where
refers to the average monthly income for household; whereas
denotes the income of bottom 40% of the population in Malaysia, amounting RM2848 per household per month (
DSM 2017). The effect of diversification in income sources on economic vulnerability had been measured as
, where
is the total number of income sources (full-time). Households with higher proportion of dependent members per gainfully employed member ratio have been estimated to appear more vulnerable (
).
3.3. Control Variables
Several variables such as gender, age (
Islam et al. 2017;
Samer et al. 2015), number of sources of income (
Al-Mamun and Mazumder 2015;
Al-Mamun et al. 2014), and gainfully employed household members (
Al-Mamun and Mazumder 2015) are discovered to affect household income and economic vulnerability. Other variables include education, household size, household income, and enterprise income. To further elaborate, in terms of gender, women’s participation in income generating activities is not common in developing countries because of the social and religious practices. Therefore, male household members with income are expected to be higher and less economically vulnerable than female. In terms of age, older participants are more skilled and experienced, they are able to enjoy higher household income compared to new participants. Additionally, participants with high level of education earn more and are less economically vulnerable than others. For gender, specifically, this study coded ‘1’ for Male while ‘0’ for Female.
4. Summary of Findings
4.1. Demographic Characteristics
Demographic characteristics of the 450 respondents, including their gender, age, marital status, government support, education, type of firm and number of years firm established presented in
Table 1.
4.2. Descriptive Analysis
In
Table 2, the mean value for the average monthly household income was RM1834.75 with the standard deviation of RM865.74. After program participation, the mean value for monthly changes in household income was RM1082.49 with the standard deviation of RM605.59. Concurrently, the mean value for economic vulnerability was 0.67% with the standard deviation of 0.59%. The number of years in development program was 10.87 years with the standard deviation of 4.43 years. Moreover, the mean value for total amount of economic loan received was RM21,454.44 with the standard deviation of RM11,167.23. The mean value for the total number of training program attended by respondents was 5.5 times with the standard deviation of 2.77 times. Their total number of training hours obtained the mean value of 40.47 h with the standard deviation of 22.87 h. The total number of centre meeting or discussion attended by the respondents was 32.77 times with the standard deviation of 20.94 times. Besides that, the mean value for age was 48.31 years old with the standard deviation of 9.612 years old. The mean value for the number of years in school was 5.82 years with the standard deviation of 3.560 years. Then, the mean value for household size was 7.80 members with the standard deviation of 1.74 members.
In
Table 3, the mean values for various groups indicated that respondents who joined the micro-credit programs for over 16 years achieved higher number of training programs and number of hours of training programs. In other words, underlying organizations might have provided useful enterprise developing training programs before. ANOVA f-test revealed that the
p-value less than 0.01, indicating that the mean value for number of training programs, number of hours of training programs and number of center meetings or discussions attended by respondents were statistically different among the groups. Furthermore, new participants were discovered to receive higher amount of economic loan as compared to existing participants because high-income households might have the ability to generate more income using their loan. Two groups of respondents after participating the programs for up to 5 years and more than 16 years had high level of ‘changes in household income’. As for economic vulnerability, the existing respondents were less economically vulnerable.
In
Table 4, the number of years of program participation, total amount of economic loan, changes in household income after the program participation, and level of economic vulnerability were grouped according to the number of training programs participated by the respondents. The mean value for this particular group showed that many existing participants received little training. The mean value for amount of economic loan showed that respondents who underwent regular trainings received more economic loans from the underlying organizations. However, there was no clear relationship between number of training received and the changes of household income, and economic vulnerability among various groups.
Table 5 presents the mean difference of number of years of participation, number of training programs attended, changes in household income after program participation, and level of economic vulnerability among various groups according to the total amount of economic loan received by the respondents. The finding revealed a relationship between economic loan received and number of years of participation, changes in household income after the program participation, and economic vulnerability. This implied that respondents received better economic loan had a higher amount of changes in household income and were economically vulnerable.
4.3. Partial Correlations
A partial correlation was performed to determine the relationship between the changes in household income, economic vulnerability and the participation indicators.
Table 6 reported that the changes of average monthly household income had a positive correlation with the number of years of participation (
p-value 0.026), number of hours spending on training programs (
p-value 0.001), and total amount of economic loan (
p-value less than 0.01) after controlling the effect of gender, age, education, and household size. In fact, economic vulnerability showed that the changes in the level of economic vulnerability had an unexpected positive correlation with the number of years of participation (
p-value 0.003). Nevertheless, a negative effect on number of training programs (
p-value 0.044), number of hours spent on training programs (
p-value 0.000), number of center meeting or discussion (
p-value less than 0.01), and total amount of economic loan received (
p-value less than 0.01) after controlling the effect of gender, age, education, and household size.
4.4. Impact on Household Income
The finding revealed the Durbin-Watson statistic of 1.920 marked the absence of auto-correlation. The VIF and tolerance values were lower than 5 and 2, respectively, indicating that no multicollinearity issue. The F and p-value from the ANOVA statistic is 4.622 and 0.000, respectively. Considering that the p-value for ANOVA is less than 0.001, it means that at least one variable can be used to model ‘changes in household income after program participation’. However, the normality of the residuals of the Kolmogorov-Smirnov test provided the p-value of 0.000, which was less than 0.05, failing to meet the assumption of normality. The Unstandardized Residual Stem-and-Leaf Plot presented the outliers based on the Unstandardized Residual values. This study removed the outliers and reanalyzed the data using 338 respondents. Since the p-value for Kolmogorov-Smirnov test of normality (N = 338) was 0.20, therefore the assumption of normality was satisfied.
Since the r2 value was 0.365, it meant that 36.5% of the variation in changes in household income after the program participation was explained by years of participation, number of training programs, number of hours spent on training programs, number of center meeting or discussion, total amount of economic loan received, gender, age, education, and household size.
As presented in
Table 7, the finding revealed that number of years of participation had a positive effect on the changes of household income after participating the development programs. Furthermore, total amount of economic loan had a positive effect on the changes of household income. The effect of number of training hours and total number of center meeting and/or discussion had a positive effect on the changes of household income. In terms of control variables, there was a positive effect of gender and education. On the contrary, there was a negative effect of age and household size on the changes of household income.
4.5. Impact on Economic Vulnerability
Given that the
r2 value for economic vulnerability was 0.545, it implied that 54.5% of the variation in economic vulnerability could be explained by number of years of participation, total amount of economic loan received, total number of training hours, number of center meeting or discussion, gender, age, education, and households size. Since the Durbin-Watson statistic was 1.078, it indicated the absence auto-correlation. In
Table 8, the VIF values for all variables were below 5, thus no multicollinearity issue was identified. The F value and
p-value of the ANOVA statistic was 58.516 and 0.000, respectively. As the
p-value was less than 0.05, at least one variable could be used to model economic vulnerability.
However, the normality of the residuals of the Kolmogorov-Smirnov test (
N = 450) provided the
p-value less than 0.01, failing to meet the assumption of normality. The Unstandardized Residual Stem-and-Leaf Plot presented the outliers based on the Unstandardized Residual values. After the outliers were removed, the data were reanalyzed using 290 respondents. The
p-value for Kolmogorov-Smirnov test of normality (
N = 290) was 0.20, thus the assumption of normality was satisfied.
Table 8 presents the standardized beta and
p-values.
With 290 respondents, the r2 value was 0.935, it indicated that 93.5% of the variation in economic vulnerability could be explained by number of years of participation, total amount of economic loan received, total number of training hours, number of center meeting or discussion, gender, age, education, and households size. Durbin-Watson statistic of 0.193 indicated the absence of auto-correlation. Considering the VIF values for all variables were below 5, there was no problem of multicollinearity identified. ANOVA analysis reported that the F value and the p-value were 450.278 and 0.000, respectively. As the p-value was less than 0.05, at least one variable could be used to model economic vulnerability.
Table 8 revealed that the effect of numbers of years of participation on economic vulnerability is negative, which indicates that participation in development programs reduced economic vulnerability among the low-income respondents. However, the effects are not statistically significant (
N = 450 and
N = 290). The number of hours of training, number of center meetings and/or discussions, and total amount of working capital showed a negative and statistically significant effect on economic vulnerability. In respect of control variables, several factors such as gender, age, and education had a negative effect on economic vulnerability.
5. Conclusions
Based on the objectives of development initiatives, this study concluded that participation in micro-credit and training programs had encouraged the changes of household income that actually decreased economic vulnerability. This finding was consistent with previous studies conducted in Malaysia (
Al-Mamun et al. 2014), India (
Ray-Bennett 2010), Bangladesh (
Schurmann and Johnston 2009), and Turkey (
Khandker 2001;
Gurses 2009). The main finding suggested that micro-credit and training programs were discovered to influence the household income and level of economic vulnerability, which heralded useful information for formulating economic and social policies as well as poverty eradication programs, especially for low-income households in Kelantan. To further augment the design of better anti-poverty policies, investigations should expand towards understanding the characteristics of households that are on average, more exposed to income shocks. This economic vulnerability could be noteworthy to deserve further examination not only for the purposes of economic performance evaluation but also its relevant sensitivity to the exposed risks of financial crisis and downturns. Additionally, it would be interesting to understand how economic vulnerability levels can be reduced, by first initiating investigation paths at the household level and then relate the exposure to risks with other variables.
Therefore, development policymakers and organizations should focus on providing flexible access to credit programs and more training and motivational programs in the areas of management and marketing support to further enhance entrepreneurial capabilities. Essentially, these training programs can yield higher micro-enterprise profits, subsequently minimizing economic vulnerability among low-income households. However, given the mandate to deliver national and regional development initiatives, development organizations should be audited by the proper authorities to safeguard against corruption and mismanagement. In addition, the beneficiaries of credit should be provided with adequate grievance and feedback channels on the performance of the respective MFIs. Finally, for the low-income household participants, micro-credit and training programs might be the key stimulus in creating small and positive impacts on overall standards of living, by way of additional income. Thus, it is necessary to establish training centres that provide enabling environments to motivate continued and frequent participation in future training programs. It would be ideal for future researches to examine the micro-credit-to-development initiatives path towards other outcomes (i.e., education, health, nutritional status), beyond just household levels, but at community and regional levels.