RE-EVALUATING SUSTAINABILITY OF MICROFINANCE INSTITUTIONS BY USING TOPSIS

Purpose: The measurement of sustainability for microfinance institutions (MFIs) has been a serious problem for both practitioners and researchers over the last few decades. A multicriteria decision-making approach is used to develop an index that measures the sustainability of microfinance institutions based on the double bottom line. Methodology: The sustainability score of MFIs operating in Pakistan for the year 2006-2015 is measured using the technique for order preference by similarity to ideal solution (TOPSIS). During the assessment, equal weights are assigned to all indicators of sustainability. Additionally, a hypothetical organization was assigned the industry threshold to generate composite scores using TOPSIS. Later, sustainability levels of individual MFIs were compared with this industry threshold. Findings: Microfinance institutions that attain higher financial sustainability and positive outreach are ranked high. The result shows that the threshold sustainability level of the microfinance sector in Pakistan from 2006-2015 was 23.52, 26.31, 23.80, 45.83, 45.83, 66.67, 77.77, 91.60, and 88.88 percent respectively. Although the sustainability level in 2015 decreases with respect to 2014, still the overall growth of the sector is remarkable. Practical implications: The results obtained from TOPSIS for evaluating the sustainability of MFIs under the double bottom line highlight its practical applicability. MFIs are under immense pressure by regulatory bodies, investors, donors, and financial experts to achieve sustainability. This index would help MFIs to track progress and improve their sustainability. Novelty/Originality: This study is the first of its kind to determine the sustainability of MFI by using all the four indicators of sustainability, including financial self-sufficiency, operational self-sufficiency, depth of outreach and breadth of outreach. Existing sustainability indicators does not provide the threshold level of sustainability. Instead, they provide a ranking of MFIs from top to bottom only. This study is novel to identify whether MFIs have met or failed to achieve sustainability by providing the threshold level. Future research determine the sustainability score of MFIs assigning to each criterion the of the may be by evaluating sustainability score across countries and comparing them. The future may the sustainability to identify other factors which affect the growth performance of MFIs.


INTRODUCTION
Sustainability of Microfinance institutions (MFIs) seeks to ensure that MFIs attain financial and operational selfsufficiency, and reach the maximum number of those deprived people of the society, attaining a 'double bottom-line', who are usually ignored by the conventional financial institutions (Hermes et al., 2011;Louis and Baesens, 2013;Rismayadi and Maemunah, 2018;Riyanti, 2018). This has led to the need for a measuring system that tracks the progress of sustainability over time (Rai and Rai, 2012;Mia et al., 2015).
The academic literature provides several pieces of evidence of attempting to develop indicators for predicting, ranking and measuring the sustainability of MFIs (Christen, 1995;Okumu, 2007;Rai and Rai, 2012). Indicators such as financial self-sufficiency, operational self-sufficiency, and outreach, measure financial and social sustainability. The subsidy dependence index (SDI) by Yaron (1994), Financial self-sufficiency (FSS) by Christen (1995), Operational selfsufficiency (OSS) by Okumu (2007), determines the most sustainable MFIs concerning a subsidy, financial, operational and outreach dimensions. The financial self-sufficiency and operational self-sufficiency have been identified as the key indicators for sustainability by CGAP (2003) and MFIs who are seeking investors' attention, proudly report their annual FSS and OSS ratios to highlight their sustainability performance.
Several researchers have criticized the applicability of existing measures of sustainability. Nanayakkara (2012), argues that the reliance on SDI is not acceptable as it indicates the dependence level of subsidies only and it may deviate MFIs from its mission of poverty alleviation as poor customers are charged with a high interest rate. Moreover, there is a strong disagreement on the financial and operational self-sufficiency measures of sustainability as these measures mainly focus on the financial aspects, ignoring the social dimension of sustainability. The same has been criticized by Bhanot et al. (2015), who stated that merely focusing on financial measures and ignoring the outreach measure deviates MFI from its original mission. There is a need to develop the sustainability measure which considers both financial and outreach measures. In an attempt to develop the sustainability index, Bhanot et al. (2015) used operational self-sufficiency and outreach measures but dropped the financial self-sufficiency, which is an important measure to determine whether MFIs can generate revenues to continue its operation without depending on subsidies. Additionally, the scope of their study was limited to only determining the sustainability scores of MFIs.
Humanities & Social Sciences Reviews eISSN: 2395-6518, Vol 7, No 2, 2019, pp 581-589 https://doi.org/10.18510/hssr.2019.7269 Lancker and Nijkamp (2000), strongly argues that only measuring the sustainability score does not indicate anything, unless it is compared with a threshold or reference value. It is useless to rank MFIs based on their sustainability scores only as there are chances that all of them may have poor performance, holding the lowest standards. Additionally, the organization with the highest rank may not necessarily be considered sustainable unless a threshold is defined. Arrow et al. (2012) also suggest that firms are considered sustainable when their performance is improved or maintained with time. This indicates that passing the minimum acceptable or threshold level is one of the conditions for firm sustainability. The lack of threshold or base value in a sustainability index is an important issue that needs to be addressed. In this paper, a threshold or base value for the sustainability of MFIs is measured by holding the view of Lancker and Nijkamp (2000), and the method proposed by Afful-Dadzie et al. (2016). Multi-criteria decision-making (MCDM) method is used in this study to develop the composite measure for the sustainability of MFIs, by using a technique for order preference by similarity to ideal solution (TOPSIS) proposed by Hwang and Yoon (1981).
In TOPSIS, alternatives are simultaneously evaluated based on benefit and cost criteria. Good performing MFI is the one which is more close to the highest score of the index and far from the lowest, whereas bad performing MFI is near to the lowest level and far from the best score (Bilbao-Terol et al., 2014; Rosli and Siong, 2018; Rotova, 2018). More specifically, TOPSIS is based on the alternative distance from two hypothetically created reference points, also called the ideal positive and ideal negative. The best alternative has the shortest distance from the ideal positive and maximum distance from the ideal negative point (Wanke et al., 2016). Additionally, TOPSIS has been widely used in industrial applications and accepted by academia (Shih et al., 2007).
Review of previous literature provides several pieces of evidence for the application of TOPSIS, but they have different perspectives compared to the current study. is similar to the TOPSIS method proposed in this paper for ranking. In this approach, sustainability scores are obtained, and MFIs are ranked accordingly. Moreover, a threshold for the sustainability of MFIs is generated to determine their sustainability level. This study is the first of its kind to determine the sustainability of MFI by using all the four indicators of sustainability, including financial self-sufficiency, operational self-sufficiency, depth of outreach and breadth of outreach. Additionally, no such study is found in the literature to determine the threshold for the sustainability of MFIs.
The remaining section of the paper is organized as follows. In the first section, the TOPSIS model is presented whereas the second section presents the method for measuring the sustainability threshold. Next, a numerical example of the TOPSIS method for MFIs is provided and is accompanied by the sustainability performance of MFIs in Pakistan for ten years. Conclusions and future research directions are then proffered.

Performance ranking using TOPSIS
Formally, the element of a decision matrix, is developed where j=1, 2….., m and k=1, 2,…, n as shown in Table1. Alternatives of the decision matrix are represented by a set of n alternatives, I={ /k=1, 2,…, n} and criteria are presented by a set of m criteria, C={ /j=1, 2,…, m}, and the relative weights associated with these criteria are presented by W= { /j=1; 2; . . .; m; =1}.
The element of the decision matrix provides the rating of kth alternative with respect to j criterion whose weight is assigned depending upon its importance. The decision matrix provides information for ranking an alternative in the matrix through the following steps; Step 1: constructing normalized decision matrix.
Firstly, a normalized decision matrix R is constructed in which dimensional criteria is transformed to a nondimensional criterion. The normalized decision matrix is measured as: Step 2: obtaining a weighted normalized decision matrix ( ) The normalized decision matrix is transformed into a weighted matrix in such a way that each criterion is assigned a weight relative to its importance. Weighted matrix is obtained when the normalized element is multiplied by its weight and is presented by equation (2) Humanities & Social Sciences Reviews eISSN: 2395-6518, Vol 7, No 2, 2019, pp 581-589 https://doi.org/10.18510/hssr.2019.7269 Step 3: determining the ideal positive ( ) and ideal negative ( ) solution.
In TOPSIS, ideal positive and ideal negative points are hypothetically generated to be used as a comparison yardstick. An alternative that is closer to ideal positive has the higher rank, and the alternative close to ideal negative has the lowest rank. Let us consider that the set of negative (less is better) and positive (more is better) criteria be denoted by , and . By using matrix R, the ideal solutions are computed as: Step 4: Use Euclidean distance for computing separation measures.
In the next step, m-dimensional Euclidean distance is used to compute the separation measures. The separation measure from ideal positive denoted by and from ideal negative denoted by are measured as: Step 5: calculating the closeness coefficient ( ) to the ideal solution.
The separation measure determines the closeness of an alternative to the ideal positive. At the same time, also determines how far the alternative is from the ideal negative. This leads to a dilemma as to which separation measure among these two is of best use. The relative closeness coefficient is used to resolve the dilemma and is presented in the equation below: The relative closeness coefficient provides a higher rank to the alternative that is close to the ideal positive point. The alternatives are ranked in descending order, from best to worst. To rank the alternatives in the ascending order, best to worst the equation (7a) can be replaced by equation (7b):

Establishing a threshold for sustainability measurement
The threshold is the minimum acceptable value below which an entity is said to have failed the sustainability test. Lancker and Nijkamp (2000), strongly argues that only measuring the sustainability score does not indicate anything, unless it is compared with a threshold or reference value. By identifying the threshold, entities are evaluated for growth performance, and it helps in the realization of effective response policies. In determining the industry standard threshold, the attributes of TOPSIS are maintained as well as competition among the alternatives is also assured.

Numerical Example
In this study, the sustainability index for MFIs in Pakistan is developed using TOPSIS-based performance ranking procedure. The sustainability score of MFIs is assessed using four different criteria, including; financial self-sufficiency, operational self-sufficiency, depth of outreach and breadth of outreach. A composite sustainability score for MFIs in Pakistan over the year from 2006-2015 is obtained using TOPSIS. Additionally, a threshold level for the sustainability of MFIs over the years is obtained.
The set of alternatives used in this study are the Microfinance institutions working in Pakistan. Column 2 in Table 2 provides a list of criteria, where C1 is financial self-sufficiency, C2 is operational self-sufficiency, C3 is the depth of outreach measured by average loan balance per borrower, and C4 is the breadth of outreach approximated by number of active borrowers, for assessing the sustainability score of MFIs. Among these four criteria's C1, C2, and C4 are designated as benefits or "more is better" criteria, which indicates that MFI is advancing towards a higher level of sustainability when FSS, OSS, and breadth of outreach (measured by the number of active borrowers) increases (Bhanot et al., 2015). On the other hand, the increase in average loan balance per borrower weakens the social role of MFIs. Smaller loan size implies that MFIs are more committed towards their social goal of poverty outreach (Cervelló-Royo et al., 2017). Therefore, C3 is designated as "less is better criteria," which indicates that as the depth of outreach (measured by average loan balance per borrower) decreases, MFI become more sustainable. Following Bhanot et al. (2015), FSS and OSS are in percent terms, while ALPB and NAB are absolute figures, so they were transformed using the natural logarithm functions. Following equation 1 the alternatives and criteria are used to develop a normalized decision matrix. In the next step, the importance of each criterion in the decision matrix is calculated by multiplying weights with the normalized decision matrix as shown in Equation (2). Column 3 of Table 2 shows the importance of weight assigned to all the four criteria of sustainability. In the current study, the sustainability of MFI is a construct based on "double bottom line" (Gutiérrez-Nieto et al., 2009), which indicates that MFIs are considered sustainable when they achieve both financial sustainability and outreach. Therefore, each criterion in the current study is assigned equal weights. Moreover The criterion weight for financial self-sufficiency, operational self-sufficiency, depth of outreach, and breadth of outreach is 0.25 respectively. The weights obtained are then used in the TOPSIS method to determine the ideal positive, and ideal negative solution based on "more is better" and "less is better criteria" by following equation (3) and (4).
Additionally, the distance from the positive ideal and negative ideal for each alternative are computed using equation (5) and (6). The relative closeness of an alternative to the ideal solution is determined, and all the alternatives are ranked accordingly. The best alternative is the one which is farthest from the negative ideal and closest to the positive ideal. The composite score is obtained by using equation (7a) which help in ranking the organizations from best to worst in the descending order. Table 3 provides the sustainability score of all MFIs in Pakistan for the year 2006-2015. The sustainability score of MFIs for each year under study and their respective rank in that year are presented in the adjacent column. It is important to note the data obtained for the study is unbalanced, and some MFIs tend not to report data over one or more years, so they were dropped for that respective year. According to Afful-Dadzie et al. (2016), the sustainability score obtained using TOPSIS lies between 0 and 1, where MFI closest to 0 tend to have lowest sustainable score and MFI having a sustainability score near or equal to 1 have the highest sustainability score. Table 3 that the sustainability score of MFIs vary across the year and their respective ranks depend on their final composite score for the respective year. ASA Pakistan has the highest sustainability score and has the highest rank for three years. Moreover, Villagers development organization has the lowest sustainability score for 2015 and is The results shown in Table 3 are merely the scores and the ranks of MFIs, and these do not determine whether MFI is sustainable or not. Lancker and Nijkamp (2000), states that "a given indicator does not say anything about sustainability unless a reference value such as a threshold is set for it." In order words, the prerequisite for sustainability is to achieve a minimum acceptable level. A threshold level needs to be defined to answer the sustainability question. When an MFI falls below this threshold level, it can be deemed to have failed a sustainability test. This is achieved by setting a standard value for all criteria which need to be fulfilled by an alternative. Moreover, Arrow et al. (2012) emphasized that firms are considered sustainable when their progress is improved with time.

It is evident from
In this study, minimum standard values for each criterion to a hypothetical alternative are assigned which help in measuring the overall relative closeness coefficient to determine the threshold level for further comparison. The minimum standard value for each criterion obtained from literature and prudential regulations for MFIs are presented in Table 4. Financial self-sufficiency ratio has a minimum standard value of 100 percent which denotes that MFIs are able to cover their cost without depending on subsidized funding and may continue on a going concern (CGAP, 2004). OSS of more than 100 percent denotes that operating revenues of MFIs exceed their operating expenses, and is a standard measure used by Bhanot et al. (2015) and CGAP (2004). According to prudential regulations of Microfinance Banks, issued by State Bank of Pakistan, the maximum exposure for both general and microenterprise loan shall not exceed PKR 500,000. Additionally, the scale parameter for number of active borrowers of an MFI shall be 10,000 and above as an indication of MFI sustainability (Bhanot et al., 2015). Operational self-sufficiency C2 100% 3 Depth of outreach (average loan balance per borrower)

C3
Maximum exposure amount shall not exceed PKR 500,000 1 4 Breadth of outreach (number of active borrowers)

C4
Greater than or equal to 10,000 Industry standard alternative value for the square of relative closeness coefficient is denoted as . For an MFI to be sustainable, it must follow the criteria mentioned in equation (8). It is important to mention here that the composite industry standard value for varies across different year. This happens in TOPSIS because the values for the ideal positive and negative solutions are derived using data for the same alternatives which changes over the years.  In addition to the above, Arrow et al. (2012) argued that sustainability might be truly represented when an entity passing the sustainability test improve its performance over time. Figure 1 provides the sustainability performance score for MFIs in Pakistan from the year 2006-2015. The sustainability threshold level for each year is measured and is presented in Table 5.  Additionally, a threshold level for the sustainability of MFI is identified by assigning minimum industry standard values of each criterion for developing a hypothetical alternative. The relative closeness coefficient of the hypothetical alternative became the threshold and was used for comparison. Due to the significance of each criterion used in this study, all were considered equally important and were assigned equal weights in sustainability measurement. This way, organizations with relative closeness coefficient below that of the industry standard alternative are deemed to have failed the sustainability test.