# An Examination of the Sustainability of Village Savings and Loans Associations (VSLAs) Amid Covid-19 and Its Impact on Household Income Levels: Lessons from Malawi, Sub-Saharan Africa | BMC Public Health

### The goal

This study aimed to assess factors associated with the sustainability of VSLAs amidst Covid-19 and their impact on household income levels.

### study design

This study used a quantitative method design to extrapolate detailed information about the subject under study [30, 31]. In this study, we used an online cross-sectional survey. The study was conducted in Malawi, Africa. Malawi is located in the southern part of Africa and borders with countries like Zambia, Tanzania and Mozambique [32, 33]. It is one of the least developed countries in the world where the majority of the population still suffers from hunger and malnutrition and lives below the poverty line [15, 34, 35]and depend primarily on agriculture for business and to feed the populace [36]. In Malawi, the study was specifically conducted in the city of Mzuzu, which is located in the northern part and the city has an estimated total population of 221,272 and a land area of ​​143.8 square kilometers (as of 2019). The study area is politically divided into 14 stations, namely; Chibanja, Chibavi West, Chibavi East, Chiputula, Jombo-Kaning’ina, Katawa, Lupaso-Nkhorongo, Luwinga, Masasa, Mchengautuwa East, Mchengautuwa West, Mzilawaingwe, Zolozolo East and Zolozolo West) as shown in Fig. 1 according to recent research Mzuzu City is one of Malawi’s fastest growing cities, largely due to its socio-economic activities, with most residents dependent on agriculture, business and working as civil servants in many of the city’s governmental and non-governmental organizations [37]. Given these conditions, it was crucial to conduct this study in this city to examine the impact of Covid-19, which could have a negative impact on people’s socio-economic status.

### data collection procedures

We have collected data from various VSLAs members operating in the city of Mzuzu. Data collection was conducted between November 2020 and January 2021. We recruited and trained a team of four research assistants, who were guided on the study’s objectives and conduct. These research assistants helped identify VSLAs groups and their members and facilitated the collection of the online survey led by Mr. Zolo (as head of research assistants). The questionnaire was sent to participants via Facebook, WhatsApp and email due to Covid-19 restrictions on gatherings, as the study was conducted when some social distancing measures were being enforced in Malawi. The questionnaire links were forwarded to the identified target groups with the support of the recruited research assistants.

### Inclusion and Exclusion Criteria

All members of VSLAs operating in Mzuzu locations who were under the age of 18 were excluded from the survey. To ensure that the respondents were from this country and city, a field was provided in the questionnaire for the respondent to indicate their country and city of residence. All those reporting outside of the study areas were excluded.

### Population, sample size and technique

The study recruited respondents who were members of VSLAs in the designated study area. A snowball and respondent-driven sampling technique was used to select the area and determine the sample population. We used this technique for the following reasons. First, data access was easy because of the researchers’ connections to individuals associated with VSLAs in the selected country. Second, due to the impact of Covid-19, it is not easy to collect the data physically, considering the social distancing measures in place in all countries [38,39,40].

Using the inclusion and exclusion criteria, we recruited 402 participants for the survey. We have a sample calculation used by Yamane with a 95% confidence level and P = 0.05 [41], N= Total Mzuzu City Population = 221,272 [37].

$$x=frac{N}{1+N{(e)}^{2}}$$

Which gave us 399 as the minimum number of participants.

### draft questionnaire

1. I.

demographic data

The first sections collected the socio-demographics of the VSLAs members, including: gender, age group, occupation, level of education, status of head of household, and number of people in the respondent’s household, which were measured in the category and coded binary ( Table 1).

2. ii.

Impact of Covid-19 on income

The second part of the questionnaire collected data on the impact of Covid-19 on participants’ income. We asked respondents to indicate the income category they earn per month before and during the Covid-19 outbreak. Income was divided into categories/groups of three income bands or levels. First were those falling below MK5,000, then those below or above MK5,000 but less than MK10,000, and finally those above MK10,000 (Table 1).

3. iii.

Performance and sustainability indicators

The third part of the questionnaire collected the performance data predicting the sustainability of the VSLAs amidst Covid-19 based on the literature. Members were asked questions about the impact of Covid-19 on; Timely loan repayments, frequency of borrowing, timely or unshared contributions, and whether members met. All variables were categorical and measured in binary form (Table 1).

### validity and reliability

A pilot study was conducted to test the instruments prior to actual data collection, involving 43 respondents including VSLAs members, Masters and Ph.D. Students. The validity and reliability of the instrument were tested by sending the instrument to experts for comment prior to the actual data collection. The research instrument was tested using Cronbach’s Alpha in SPSS and found to be 0.8.

### Ethical Approval

The ethical acceptability of this study was reviewed and approved by the School of Economics and Management of Yangtze University (permission number REF/YU/2020/08 (Fig. 3)) and the Mzuzu City Council (permission letter reference number MCC/dated August 12), 2020 (Fig. 4)). In addition, researchers respected and followed the 1964 Helsinki Declaration when conducting research involving humans. Participation in the survey was voluntary and participants gave their informed consent to this questionnaire before completing it.

### data analysis

After collecting the data using the Google form, we coded it in Microsoft Excel and later imported it into SPSS version 23 for analysis. We presented the descriptive statistics results using frequency tables, graphs and charts. The chi-square test was performed to determine associations between sociodemographic variables and other variables. P valuewas statistically significant p

### Econometric model specification

This was used to answer our second research question: to predict the factors associated with the sustainability of VSLAs. We used Future of VSLAs as our dependent variable, which was coded or characterized as a two-category variable. The coding was that when the VSLAs members confirm that they have certainty about the future and sustainability of VSLAs in relation to the current situation of Covid-19, the value reported was 1; otherwise 0. Therefore, the dependent variable dichotomy instructed and proposed to us used the binary logistic regression model deemed appropriate by other scholars[14, 42, 43].

In this study, a logistic regression model, the dichotomous variable is defined as:

$$y={int }_{0}^{1}while 1=The presence of the features 0=The absence of the features$$

while the odd is defined as,

$$Odds=frac{p}{p-1}=frac{The probability of the presence of the features}{The probability of the absence of the features}$$

while the definition of the logit model is such that

(Logitleft(pright)={beta }_{0}+{beta }_{1}{X}_{1}+{beta }_{2}{X}_{2 }+{beta }_{3}{X}_{3}+cdots +{beta }_{k}{X}_{k})[43, 44]

Where p represents the probability of the presence of the feature of interest. The logit transformation is defined as the log of odds.

(mathrm{log}left(frac{p}{1-p}right)=Logitleft(pright)={beta }_{0}+{beta }_{1} {X}_{1}+{beta }_{2}{X}_{2}+{beta }_{3}{X}_{3}+dots +{beta }_{k {X}_{k}+e)[43, 44]

Whereas;

({beta}_{0})= constant,

({beta}_{1}-{beta}_{k})= are the coefficients of logistic regression,

({X}_{1}-{X}_{k})= are independent explanatory variables, and.

(e)= is an error term.