Content uploaded by Roy Mersland
Author content
All content in this area was uploaded by Roy Mersland on Feb 16, 2021
Content may be subject to copyright.
Is Employee- Client Matching Good for Firms Targeting the Bottom of the Pyramid? A
Study of Microfinance Institutions
Naome Otiti*, Kjetil Andersson & Roy Mersland
School of Business and Law, University of Agder
Gimlemoen 19, 4630, Kristiansand, Norway
*Corresponding Author: naome.otiti@uia.no
Accepted for Publication in International Journal of Development Issues
DOI: https://doi.org/10.1108/IJDI-04-2020-0069
Abstract
Purpose: The purpose of the study is to determine whether there exists employee-client
matching at the Bottom of the Pyramid (BOP) as well as the most favourable employee-
client categorization in terms of employee productivity when serving the BOP market. This
is important in a bid to determine how to effectively operate at the BOP given the market’s
unique characteristics.
Design/methodology/approach: This study uses two methods depending on the research
question. Firstly, a one-way analysis of variance (ANOVA) is used to determine the
different employee-client categories based on socio-economic status. Secondly, Fixed
Effects analysis are performed based on these categories to determine the most suitable
employee-client category.
Findings: The results show the existence of employee-client matching based on similar
socio-economic status. However, multivariate testing reveals that the mismatch category,
where employees are of higher socio-economic status than the clients, generates more
favourable employee productivity. Moreover, this result may be contingent on the
geographical location of the firm.
Practical implications: The findings are important for human resource management
particularly, the employment strategy of BOP firms. It suggests the need to consider
employee profiles as well as client profiles when deciding which new markets to target.
Originality/value: The paper uses a global database of microfinance institutions as a case
of BOP firms to investigate employee-client matching at the bottom of the pyramid.
Key words: Employee-Client Matching, Socio-economic Status, Homophily, Bottom of
the Pyramid, Employee Productivity, Microfinance
1
1. Introduction
Who should a firm hire when targeting clients at the bottom of the pyramid? Should it hire
employees from the same social strata as its clients or would it be beneficial to have employees
from higher or lower social strata than its clients? We address these questions in this paper.
Over the past couple of decades, the bottom of the pyramid (BOP) market has received significant
recognition. It is the largest market in terms of number of potential clients consisting of the
approximately 4 billion of the world population (Prahalad & Hart, 2002; Hammond et al., 2007;
Schuster & Holtbrügge, 2012; Bocken, Fil & Prabhu, 2016). Within the BOP market, some clients
are considered less poor than others thus, diversifying the clientele. Moreover, there are enormous
opportunities for firms due to the aggregate demand embedded within this market (Prahalad,
2006). Despite this, a common challenge for firms targeting the BOP is how to effectively deliver
products and services.
Firms operating in the BOP market include social enterprises and non-profits as well as regular
for-profit corporations. Microfinance institutions (MFIs), which are the subject of this study, are
social enterprises operating with both financial and social logics (Battilana & Dorado, 2010).
Microfinance institutions started out in the 1980s as socially oriented institutions with the goal of
lending to the unbanked (Morduch, 1999). Since then, the microfinance industry has grown
tremendously and evolved to include institutions ranging from non-profit entities to commercial
banks.
Hart & London (2005) suggest that in order to effectively serve the BOP market, firms should
become ‘indigenous’ to the society. In other words, firms should become embedded within the
market segment and ultimately gain client acceptability. This suggests that serving these markets
is not merely about providing affordable low-cost products. A step in becoming indigenous may
also involve hiring employees from the local target community as they are believed to have a better
understanding of the clients’ beliefs and practices. They also have the added advantage of being
able to communicate in the local language unlike employees hired from outside the market
(Kennedy, 2012; Banthia et al., 2011).
Additionally, the social characteristics of employees and clients may influence operations at the
BOP (Labie, Méon, Mersland & Szafarz, 2015). For instance, Ahmad (2002), in his study of non-
2
profit firms in Bangladesh, suggests that gender similarities between employees and clients
facilitate the provision of non-financial and financial services to women. Beck, Behr & Madestam
(2011) find that there is less likelihood that first-time clients will return for another loan when
served by employees of the opposite gender. This preference for similar others may apply not only
to clients but also to employees, who may prefer association with clients who are similar to them
(Labie et al., 2015). Along these lines, international development organizations like the World
Bank require their partners to hire local employees in a bid to establish strong internal ties with
clients in the countries of operation (Kennedy, 2012). Overall, this literature indicates homophilic
tendencies (McPherson, Smith-Lovin & Cook, 2001) at the BOP, suggesting the benefits of
employee-client matching a concept that has been explored mainly in the upper levels of the
economic pyramid (for example, Gonzalez, 2013; Avery et al., 2012).
In this paper, we seek to answer two research questions. The first is whether employees and clients
of BOP firms are matched based on similar socio-economic status. The second is whether
employee-client matching on socio-economic basis enhances employee productivity when firms
target BOP markets. Our theoretical foundation is the similarity attractiveness paradigm (Byrne,
1971). It suggests that individuals tend to prefer association with those with whom they share
similarities along certain social dimensions. Thus, in our study we assume that clients prefer
employees similar to them and, vice versa, employees prefer clients similar to them.
Oakes & Rossi (2003) consider socio-economic status as a measure of differences in access to
necessary resources. Other scholars refer to it as a measure of individual economic and social
differences in terms of income, occupation, or education (Adler & Snibbe, 2003). Since some MFIs
focus on the very poor whereas others focus on the less poor (Labie et al., 2015; Armendariz &
Szafarz, 2011), the socio-economic status of employees and clients presents a potential dimension
along which to investigate employee-client matching at the BOP. Additionally, the interpersonal
nature of the employee-client relationship in MFIs (Siwale & Ritchie, 2012) presents a solid
foundation for studying this matching concept.
Empirical evidence outside the BOP literature finds that matches between employees and clients
are characterized by improved performance and productivity outcomes particularly in the sales
field and more recently in banking (Gonzalez, 2013; Fisman, Paravisini & Vig, 2017).
Nevertheless, some scholars suggest that although individuals may prefer association with others
3
that are similar to them, matching may not have a positive impact on performance. For instance,
Dwyer, Orlando & Shepherd (1998) suggest that matching may demotivate employees and limit
their sales potential since they are restricted to a particular clientele. These mixed findings on non-
BOP markets make the logics behind matching employees and clients uncertain. In the middle
ground, Joshi & Roh (2009) suggest that context is an important determinant of the outcomes of
employee-client matching. The unique characteristics of the BOP relative to other market segments
therefore make it an interesting case to study employee-client matching.
We use data from 474 MFIs operating in 87 countries. The employees and clients’ socio-economic
status are measured by average salary and average loan size respectively. Regarding the first
research question, we find very strong support for the existence of matches between employees
and clients based on similar socio-economic status. For the second research question the
association between employee-client matching and employee productivity – the results do not
support our assumption of a more favourable outcome in the presence of socio-economic
similarity. We find that employee productivity is most favourable when employees are of higher
socio-economic status than the clients. Our results also show that the least favourable employee-
client category is that where lower socio-economic status employees serve higher socio-economic
status clients. This seems to suggest that higher socio-economic status employees are more
favourable regardless of whether they are matched with very poor or less poor clients. Thus, the
beneficial effect of similarity matching on employee productivity is supported only for the high-
status employee/high-status client category.
Our study responds to the call for more human resource research in BOP enterprises (Labie et al.,
2015). One take-away from this study is that in terms of productivity, MFIs should not match low
status employees with high status clients. Whether this result generalises to other BOP firms which
depend on the employee-client relationship, is a topic for future research. Regarding the concept
of local engagement with the community through hiring locals, our results suggests that this will
not be beneficial in terms of productivity. Furthermore, our results hint that commercialization
may lead to less engagement with the community in terms of hiring locals. Our results present a
puzzle with regard to similarity-attractiveness hypothesis. We find that firms actually match
employees and clients based on similarity, however, the rationale for this remains unclear.
4
Ultimately, it may be important to consider regional variations when establishing the employment
strategy of such firms.
The paper proceeds as follows. In the next section, we present a background to the
commercialization aspects in microfinance. In Section 3, we present our theoretical framework
and develop testable hypotheses. In Section 4 we present the data and methods. In Section 5 we
present the results. In Section 6 we discuss our results and Section 7 concludes.
2. Background: Aspects of Commercialization in Microfinance
Social enterprises like microfinance institutions are firms that combine a social and financial logic.
Therefore, due to this dual logic, they can be described as hybrid firms (Doherty, Haugh & Lyon,
2014; Battilana & Dorado, 2010). These firms mainly seek to fulfill the needs of society that
remain unmet by the public sector and private organizations. Thus, the social mission is considered
as core and tends to vary among social enterprise types. Common examples of social missions
include poverty alleviation, employment creation, inequality reduction and environmental
protection (Doherty et al., 2014). As such, microfinance institutions like other social enterprises,
can be viewed as firms that contribute to social change at the bottom of the pyramid markets (Goyal
et al., 2014). To remain sustainable, it is important for such hybrid firms to balance the social and
financial objectives.
However, balancing these objectives may be a challenge for such firms due to the competing nature
of the institutional logics that they ascribe to. Indeed, there is vast discourse in social enterprise
literature with terms such as ‘trade-offs’ and ‘tensions’ being used to reflect this complexity. For
instance, Wry & Zhao (2018) find that microfinance institutions experience a trade-off when there
is more focus on the social objective relative to the financial objective. Likewise, since these firms
tend to have various stakeholders, pressures may arise from them due to their differing logics
leading to tensions in the firm. Battilana & Dorado (2010) offer a good illustration of this in their
study of a Latin American MFI, they found that there were intergroup tensions among the
employees with different ideologies. Recent studies in other BOP markets also find that
partnerships between different institutional types for example, between banks and MFIs, may
result in conflicts due to their competing logics (Parekh & Ashta, 2018). These types of tensions
5
bear resemblance to Smith, Gonin & Besharov ’s (2013) conceptualization of belonging tensions
which raises questions about identity in the social enterprise. For instance, questions about “who
we are” and “what do we do” are raised since employees may identify more with either social or
financial logic. However, Battilana et al., (2015) suggest that a solution to this may involve the
assignment of responsibility for the social or financial activities to different employee groups. This
could also apply in microfinance institutions with their diverse client groups and equally diverse
employees based on socio-economic status. Thus, it could be beneficial to match employees and
clients based on socio-economic status.
Furthermore, the commercialization of microfinance characterized by investor take-over of the
microfinance industry has led MFIs to behave as banks (Ledgerwood & White, 2006). It has led
to the increased demand for financial returns as investors and shareholders become residual
claimants on the profits. A popular case is the Mexican MFI Compartamos which transformed to
a bank and by launching an IPO made its early investors very wealthy (Mersland & Strøm, 2010;
Ledgerwood & White, 2006). However, this raised some criticism as it unveiled the atrocious
interest rates of almost 100 percent charged to its poor clients (Cull, Demirgüç-Kunt & Morduch,
2009). Moreover, increased commercialization prompted some formal banks to downscale and
include microfinance products and services (Bell, Harper & Mandivenga, 2002). In their study on
the impact of commercialization on MFI operations, D’Espallier et al. (2017) found an increase in
average loan size following transformation from NGO to bank. This therefore suggests an
upscaling in the clientele from the very poor to less poor. Some might argue that this outcome of
commercialization could negate the fact that all microfinance institutions are social enterprises.
Furthermore, the proliferation of commercialization in microfinance has influenced some firms to
prefer the financial goal over the social goal suggesting mission drift (Mersland & Strøm, 2010;
Copestake, 2007). This implies that some MFIs tend to favour richer clients as opposed to poorer
clients. It has been the subject of much debate in microfinance since it implies deviation from
microfinance’s social goal of extending loans to the unbanked poor (Hermes, Lensink & Meesters,
2011). Nevertheless, others argue that through serving rich clients, MFIs are able to provide loans
to poorer clients (Mersland & Strøm, 2010), moreover, at lower interest rates. Additionally,
mission drift may also result from employee activities in the field. Beisland, D’Espallier &
Mersland (2019) find evidence of this where the employees are less likely to serve vulnerable,
6
hence poorer clients. This phenomenon referred to as personal mission drift (Beisland et al., 2019).
Other studies rather suggest that there is a tendency for employees to favour clients with similar
social aspects to them, suggesting a form of discriminatory behaviour in employee interactions
with clients (Labie et al., 2015).
3. Theoretical Framework
Similarity Attractiveness Paradigm: Homophily Effect
According to Byrne (1971), similarity-attraction can be said to occur when individuals in society
seek association with groups or individuals with whom they share similarities. A similarity is the
extent to which members of a group share personal or other social characteristics (Smith, 1998).
Also referred to as the homophily effect, similarity-attractiveness is based on different social
dimensions, such as age, gender, socio-economic status, and religion (McPherson, Smith-Lovin &
Cook, 2001). When noticeably different groups exist, individuals tend to perceive members of
their in-group as being similar to them and members of the outgroup as being dissimilar. The fact
that individuals belong to the same group can thus be viewed as creating a sense of belonging,
which in turn encourages cooperation. Therefore, similarity attractiveness illustrates how different
social factors influence the behaviour of individuals.
Studies in different fields have shown that individuals tend to be attracted to those who are similar
to them. For instance, donors and lenders have been found to prefer funding individuals with whom
they can relate (Loeweinstein & Small, 2007; Galak, Small & Stephen, 2011). In the sales field,
Dwyer et al. (1998) find that sales personnel tend to prefer potential clients with whom they share
similarities in gender or age. It has been suggested that similarities influence trust, communication
and satisfaction (Rai et al., 2009; Smith, 1998; Byrne,1969). Therefore, in a firm setting, this
preference for similar others may have an important influence on the relationship between
employees and clients and, hence, on the performance of the firm.
In the BOP field, the tendency to prefer similar others has also been demonstrated. The unique
characteristics of clients may require BOP firms to hire employees from the local community. For
instance, in an initiative to extend financial services to people in rural areas of Uganda, local
employees were found more suitable than non-locals due to their knowledge of the clients’ beliefs
and practices (Banthia et al., 2011). As part of this initiative, an employee-client gender pairing
was proposed for activities that require face-to-face interaction. In another study, female clients in
7
South Asia were found to prefer fellow female employees when receiving non-financial services
(Ahmad, 2002). In addition, van den Berg, Lensink & Servin (2015) in their study of a Mexican
MFI, found that the employees hired were usually of the same Catholic religion as the majority of
their potential clients. This similarity facilitated communication and commitment to the MFI.
BOP firms do not serve a uniform socio-economic segment. Some focus more on ‘lower poor’
clients whereas others target more ‘upper poor’ clients (Labie et al., 2015; Armendariz & Szafarz,
2011). Besides that, MFI employees also constitute different socio-economic statuses (Siwale,
2016). Thus, socio-economic status presents a potential social dimension that can also impact the
employee-client relationship in the BOP market since it is not necessarily homogeneous. For
instance, in microfinance, it is reported that some employees tend to look down on poorer
customers and develop pretentious attitudes when dealing with them (Jacobs & Franceys, 2008).
Other studies find that employees tend to favour urban clients, who are usually less poor than rural
ones (Beisland et al., 2019; Labie et al., 2015).
More generally, the aspect of socio-economic status has been widely explored in literature,
moreover from different perspectives. From a cultural perspective, it is viewed as an environment
in which individuals are embedded which shapes how they perceive themselves as well as the
nature of their relations with others (e.g., Stephens, Markus & Fryburg, 2012). Likewise, some
studies focus on the consequences of belonging to a particular socio-economic group (Shah,
Mullainathan & Shafir, 2012). Other scholars rather conceptualize socio-economic status as a
rank-related construct (Kraus, Tan & Tannenbaum, 2013). Under this, comparisons of one’s
resources determine their rank in society relative to others. Thus, when individuals of different
socio-economic status encounter each other, cross-class encounters are said to occur and
individuals tend to respond to these differences in interpersonal and intrapersonal ways (Gray &
Kish-Gephart, 2013).
Considering socio-economic similarities, the study of Byrne, Clore & Worchel (1966) was among
the first, albeit in a student setting. Using one’s likelihood of enjoying working with another as a
measure of attraction, they found that individuals of similar socio-economic status were more
attracted to each other than to dissimilar individuals. Related studies in the health sector also show
that doctors prefer to treat high socio-economic status patients, who presumably are similar to
them (Willems et al., 2005). This preference is attributed to the greater ease of communication that
8
facilitates information exchange. Although the study seems to characterize all employees (doctors)
as being of higher socio-economic status, it is nevertheless relevant to our study as it illustrates
individual preferences for others of similar socio-economic status.
Employee-Client Matching
Based on the above discussion, the homophily effect can arise for two (not mutually exclusive)
reasons. First, it could be that employees seek clients that are similar to themselves. Secondly, it
could be that clients seek firms with employees that are similar to themselves. In this study we
cannot distinguish between who seeks whom or not.
Employees can be categorised into two types and clients into two types. A firm could hire
employees with high socio-economic status or low socio-economic status1. It could also serve
clients with high socio-economic status or low socio-economic status. This therefore gives us four
mutually exclusive employee-client categories, as shown in Table 1.
Regarding our categorization in Table 1, we cannot say in absolute terms, for instance, that in the
HL category, employees are of higher status than clients. Rather, we mean that this is the case
relative to the LL category. Put simply, this implies that the social distance between employees
and clients is higher in HL category than in LL category. The same explanation applies for the LH
category. Nevertheless, for the rest of the paper, we present as though the mismatch categories are
in absolute terms.
[Insert Table 1 Here]
If the homophily effect is supported, we would expect to find more firms in quadrant (1) and
quadrant (4). Thus, our first hypothesis is simply related to whether there is matching or not at the
firm level.
Hypothesis 1: There is matching between employees and clients for MFIs based on similar
socio-economic status.
We might also observe matching where firms instruct their employees to match with similar
clients. Presumably, firms would do this if there were a beneficial effect of matching on
performance outcomes, which brings us to our second hypothesis.
9
Employee-Client Matching and Its Impact on Performance
Employee-client matching is considered by some scholars to be a suitable management strategy
for firms to achieve a competitive advantage (Morrison, 1992; Cox & Blake, 1991). The
performance benefits of employee-client matching are mainly attributed to trust and more open
communication between similar individuals (Avery et al., 2012). Most empirical studies on
matching and performance have focused on demographic factors, like age (Dwyer et al., 1998),
gender (Kochan et al., 2003; Dwyer et al., 1998), and race (Gonzalez, 2013; Avery et al., 2012;
Kochan et al., 2003). We abstract from these studies and investigate the relationship between
matching based on socio-economic status (as illustrated in Table 1), and employee productivity.
Empirical evidence shows that certain firms mainly in the sales and marketing field tend to
incorporate a matching strategy. In a study on retail store productivity, Avery et al. (2012) find
that racial-ethnic matches between employees and clients improve customer satisfaction and hence
positively influence productivity. Another related study finds that employee-client matching based
on race has a stronger influence on a firm’s financial performance in culturally diverse
communities compared to homogeneous ones (Gonzales, 2013).
In a study on gender and age matching, Dwyer et al. (1998) find that sales employees are attracted
to clients that are similar to them. However, on closer analysis, it was found that age similarities
have no significant impact on performance (productivity), whereas gender mismatches have a
positive one (Dwyer et al.,1998). These mixed findings suggest that matches do not always have
a positive influence on firm performance, that is, employee productivity.
Furthermore, aspects of employee-client matching also seem to be exhibited in areas dominated
by strong cultural beliefs and practices. In a study of an Indian bank, Fisman et al. (2017) find that
when a banking officer and a client belong to the same cultural group, there is increased credit
access, repayment, and dispersion of loan size compared to when there are cultural differences.
They attribute such performance outcomes to greater ease in communication as well as the banking
officer’s ability to issue social sanctions on a culturally similar client.
The BOP literature that suggests employee-client matching also hints at its potential benefits. For
instance, van den Berg et al. (2015) find default rates in MFIs to be lower when an employee is of
the same gender or religion as the client. Beck et al. (2011) find that there is less likelihood that
10
first-time clients will return for another loan when served by employees of the opposite gender.
These studies suggest that matches based on other social dimensions such as socio-economic status
are also likely to influence performance outcomes. Thus, focusing on employee productivity as
our performance outcome of interest, we propose the second hypothesis of our study.
Hypothesis 2: MFIs with similar employee-client socio-economic matching have higher
employee productivity than MFIs with dissimilar employee-client matching.
Put differently, firms in quadrants (1) and (4) in Table 1 are predicted to have more favourable
employee productivity because of the similar socio-economic match between employees and
clients.
4. Data and Methods
Data and Sample
We use data from a secondary dataset extracted from compilations of risk assessment reports of
five specialized microfinance rating agencies namely, Microfinanza, MicroRate, Planet Rating,
CRISIL and M-CRIL. These rating agencies are internationally recognized and originally
approved by the Consultative Group to Assist the Poor (CGAP), the microfinance branch of the
World Bank. The rating reports consist of information about the MFI’s governance, management,
and financial and social operations. Appendix 1 shows the distribution of MFIs by country and
region in our dataset. The final unbalanced panel sample consists of 474 MFIs for the period 1998
- 2015 with data on 1923 observations per variable used. Rated data is not randomly drawn from
the population of MFIs, therefore there is a risk of sample selection bias. Nevertheless, rated data
and the current dataset have been used in many influential MFI studies. For example, Garmaise &
Natividad (2010) use it to examine the influence of asymmetric information on the financial and
operating activities of MFIs, Gutiérrez-Nieto & Serrano-Cinca (2007) use it to determine the
factors affecting the rating of MFIs, whereas Randøy, Strøm & Mersland (2015) use it to show
how entrepreneur CEOs affect MFIs. Moreover, increasing access to external funding is one of the
motivations to attain a microfinance rating (Mersland & Strøm, 2010). This exposes rated MFIs to
investors with either social or financial orientation. Therefore, our dataset includes MFIs with both
very poor and less poor clients, this diversity in clientele being relevant for our study. Additionally,
11
compared to other data sources which are self-reported, rated data is reported by a third party
representing its authenticity2.
Matching Variables
To answer research question one, we need to create a set of variables that measure whether a firm
belongs to quadrant (1), (2), (3) or (4) in Table 1. Also, these variables will enter as independent
variables for research question two.
We use average loan size per client to proxy for the socio-economic status of the clients. This
proxy is commonly applied in microfinance research to indicate the socio-economic segment that
an MFI targets (Cull, Demirguz-Kunt & Morduch, 2007; Schreiner, 2002; Mersland & Strøm,
2010). Similarly, we use average salary per employee to proxy for the socio-economic status of
the employees. As in other studies (for example, Abate, Borzaga & Getnet, 2014; Périlleux, Hudon
& Bloy, 2012), this proxy is obtained by dividing personnel costs by the number of employees.
The firm has the following employee-client choices. It can hire employees of high socio-economic
status or low socio-economic status, that is, high or low salary employees respectively. The
assumption is that employees in the first group receive a higher salary than employees in the second
group simply because of their higher socio-economic status or because they have higher education
and/or experience. Additionally, the firm can choose to serve clients of high socio-economic status,
or low socio-economic status, that is, high or low average loan size clients, respectively. The
assumption is that clients in the first group receive a higher average loan size than clients in the
second group (Cull et al., 2007; Schreiner, 2002).
The firms can now be classified into four mutually exclusive employee-client categories as
indicated in Table 1: high-status match, low-status match, employees serving downwards and
employees serving upwards. In our analysis, these are denoted by , LL, HL, and LH
respectively3. To distinguish between high and low socio-economic status we apply yearly country
median values of loan size and salary, that is, median values for each year per country where the
MFI operates. Similar research by Beck et al. (2011) also uses median values to distinguish
between high and low social distance between MFI employees and clients.
12
Thus, let ( denote average salary (average loan size) of employees (clients)
from MFI in country j in year , and let
(
) denote the median value of these
variables in country in year . The following dummy variables define the 4 categories:
(1)
A firm is categorized as (high-status match) in year if its employees’ salary in year is
higher than the median salary of employees from the MFIs in the country in year , and its clients’
loan size in year is higher than the median loan size of MFI clients in the country in year , and
so forth.
The Impact of Matching on Employee Productivity
To answer research question two on whether employee-client matching has an influence on MFI
employee productivity, we use the variables defined above in the following fixed effects
regressions:
= + + + + + ( 2)
Greek letters denote coefficients to be estimated, are different control variables (see
Table 2 below), and is an error term assumed to be independent and identically distributed
(). The matching coefficients of , , and are included; hence, the s are the marginal
effects relative to the left-out category, , that is, employees serving upwards.
In the above regression, our performance measure is defined as the
number of clients served per MFI employee (MicroRate, 2014; Hudon & Traca, 2011). According
to the MicroRate (2014) report, employee productivity is viewed as providing an institution-wide
perspective on an MFI’s performance as it considers all the MFI employees. Furthermore, since
MFIs seek to expand their client base, employee productivity uniquely distinguishes MFIs from
traditional banks, which aim to increase their portfolio size (MicroRate, 2014).
13
Control Variables
To control for macroeconomic institutional variables, the socio-economic statuses of the
employees and clients are scaled by PPP-adjusted gross domestic product per capita (GDP per
capita, PPP $).
Controls for MFI-specific variables are also considered. We control for MFI size and MFI age
since bigger and older firms are expected to perform better than smaller and younger firms. Bigger
firms perform better because of economies of scale, as has been confirmed in the microfinance
industry (Hartarska, Shen & Mersland, 2013). Older firms perform better because of learning
effects (Zamore, 2018). Size is proxied by the natural logarithm of a firm’s total assets and age is
the number of years that the firm has been in operation. Furthermore, the market of operation has
the potential to influence employee productivity since rural clients are generally more dispersed
and poorer than urban clients (Cull et al., 2009). Thus, a dummy variable indicating whether or
not the MFI has some operations in urban markets is included. The credit methodology of an MFI
is likely to impact productivity and is denoted by a dummy variable indicating whether the main
lending method is individual lending or group lending. We also control for the credit risk of the
MFI. It is represented by a combined measure of PaR30 (likelihood of loss in the future) and a
write-off ratio (unrecoverable loans written off), since these two variables together reflect a truer
credit risk of an MFI than either of the two variables alone (Lensink, Mersland, Vu & Zamore,
2018). Finally, since the type of ownership of MFIs is not universal, we control for whether the
MFI is incorporated as a shareholder or non-shareholder- owned firm. Consistent with other
studies, we include this variable based on the fact that the type of ownership may account for
variations in performance across firms (Williams & Nguyen, 2005).
A summary of the variables described above is presented in Table 2.
[Insert Table 2 Here]
Table 2 presents means, standard deviations and the description of variables used to answer the
two research questions. All monetary variables are dollarized and real in the sense that they are
deflated by nominal GDP per capita and PPP-adjusted. Thus, for example, the mean value of salary
reflects that the average annual salary of an employee in our sample is 13 per cent higher than the
14
PPP-adjusted GDP per capita in the country. On the other hand, the mean value of loan size is 17
per cent of the GDP per capita PPP-adjusted in the country. The descriptive statistics on the
original data showed some unreasonable figures, therefore, the data has been winsorized at the 1
per cent and 99 per cent cut-off levels. Using the original data does not change the qualitative
results.
5. Results
Research Question 1: Existence of Socio-economic Matching
To determine whether MFIs match their employees and clients (research question 1), we
performed a one-way analysis of variance. Table 3 below displays the result of the test. For
reference we also repeated the means of the matching variables from Table 2. Specifically, we let
these means (frequencies) be denoted by . In the case of no matching of employees and client of
similar socio-economic status, we expect the distribution of , , , to be random, with
= 0.25 for each category. In the presence of such matching, however, we expect the frequency of
and to be significantly higher than 0.25.
[Insert Table 3 Here]
From the table, we see that about 58 per cent of the sample is a similar socio-economic match, that
is, or . Moreover, we find that the frequency µ is significantly different from 0.25 at every
conventional significance level. We conclude that the data support Hypothesis 1, that there is
matching of employees and clients based on similar socio-economic status. This result is very
robust to alternative classifications. For instance, if the categories comparing salary and
were instead defined relative to their yearly world medians, that is, the median values in each year
based on all countries in the dataset, then the sorting according to similar socio-economic status
result would remain intact. Therefore, the data strongly corroborates that there is similar socio-
economic employee-client matching in MFIs.
15
Research Question 2: Socio-economic Matching and Its Impact on Employee Productivity
Having established that there is a clear tendency for employee-client matching based on similar
socio-economic status, we turn to the immediate question of how it impacts employee productivity.
Columns (1) and (2) of Table 4 report the fixed-effect results for equation (2) using employee
productivity as the dependent variable4. As can be seen, column (1) shows the results without the
control variables, while column (2) shows the results with control variables. We include a country-
specific time trend in all regressions. Initial regressions show that error terms are heteroskedastic;
therefore, cluster-robust standard errors are reported, as per Arellano (1987).
[Insert Table 4 Here]
Comparing the models with and without control variables, we see that the qualitative results are
the same with respect to our variables of interest, that is, the category dummies. We observe that
loan size, urban market, MFI size and risk are significant and the R-squared is higher in the
regression with controls. Interestingly, salary (the control variable) is not significantly associated
with productivity.
Regarding our variables of interest, that is, HH, LL and HL, recall that all coefficients must be
interpreted in relation to the left-out category, that is, LH: employees serving upwards. Regarding
regressions (1) and (2), recall that if Hypotheses 2 is to be supported then the coefficients on HH
and LL must be positive, and each must be larger than HL. In regressions (1) and (2) we see that
the coefficients on HH and LL are significantly larger than the coefficient on LH; however, the
coefficient on HL is significantly larger than the coefficients on LL and HH. Thus, we do not find
support for the hypothesis that similar employee-client matching based on socio-economic status
is associated with higher performance. On the contrary, dissimilar matching of the HL type, where
high-status employees are matched to low-status clients, is the best type of match in terms of
employee productivity. Put differently, having low-status employees serve high-status clients is
the least favourable match for employee productivity whereas having high-status employees serve
low-status clients is the most favourable match.
Robustness Checks
The different employee-client classifications may have different meanings in different
geographical locations due to culture differences in dimensions like power distance and socio-
16
economic differences. Power distance measures individual behaviour and regard for those in
positions of authority (Hofstede, Hofstede & Minkov, 2005; Hofstede, 2001). High power distance
suggests that inequalities are more normalized in these societies such that individuals at lower
levels of the hierarchy tend to look up to those in higher levels (Hofstede et al., 2005; Hofstede,
2001). This suggests that our matching variables may have different meaning in different cultural
settings. We expect the cultural differences to be much less within more narrowly defined regions
compared to the whole sample. Moreover, differences across regions are of interest in itself. Thus,
we split our sample into 5 regions based on the World bank’s categorization, that is, Latin America
and the Caribbean (LAC) , Sub-Saharan Africa (SSA), Europe & Central Asia (ECA), South-East
Asia & the Pacific and Middle East (SEAP) and North Africa (MENA). Table 5 reports the results
of separate regional productivity regression analyses.
[Insert Table 5 Here]
From our robustness checks in Table 5, we find that in Latin America & Caribbean as well as
South East Asia & Pacific regions, qualitatively, our results from the full sample in Table 4 are
maintained. However, we observe that coefficients of the categorical variables are higher in the
South East Asia & Pacific region. Specifically, we find that in these regions relative to our
reference category (LH), the best performing category variable is HL. For Europe & Central Asia,
Sub-Saharan Africa and Middle East and North Africa, no category variables are significant at the
5% significance level.
We suggest possible reasons for our results in the next section.
6. Discussion
The results of the study provide strong support for Hypothesis 1, which predicts that socio-
economic similarity matches between employees and clients of MFIs are more likely to occur in
BOP markets. In line with the homophily effect, our results can be attributed to the behavioural
tendency of individuals to prefer others that are similar to them (Byrne,1971; McPherson et al.,
2001). Previous studies suggest that it is easier for individuals to relate to similar others since it
facilitates communication and the development of trust (Rai et al., 2009; Jones, Moore, Stanaland
17
& Wyatt, 1998). Thus, similarities between individuals can be considered important in establishing
an employee-client relationship.
In this study, the basis for our argument is an individual-level theory with the expectation that
benefits associated with employee-client similarities will be reflected in employee productivity.
However, our results are puzzling as the matching categories do not provide support for higher
productivity outcomes where there are employee-client similarities relative to dissimilarities. The
results seem to suggest that the firm has no strong incentive to engage in similarity matching
particularly if it can choose its clients and employees freely. If an MFI targets poorer clients,
employee productivity will be higher if the firm hires high-status employees. For high-status
clients, employees of similar high status are preferable. Thus, high-status employees appear to be
the most preferable regardless of the clients’ status.
Employees of higher status tend to be more educated (Siwale, 2016; Adler & Snibbe, 2003) and
hence more skilled in problem solving, communication, and interpersonal relations (Bruns,
Holland, Shepherd & Wiklund, 2008). These skills are likely to facilitate employees’ interactions
with potential clients and hence increase productivity. Moreover, it is generally accepted that
individuals from low income segments tend to have less education levels and fewer skills than
individuals from higher income segments (SadreGhazi & Duysters, 2009; Chhibber & Nayyar,
2008; Prahalad & Hart, 2002). This may explain why the match categories with high-status
employees (that is, HL and HH) were seen to outperform those with low-status employees (that is,
LL and LH). Moreover, irrespective of the client type that the MFI is targeting, in terms of
productivity, our results suggest that it is better to have high status employees. For example, this
implies that even MFIs required to meet a predetermined number of clients or serve as many poor
clients as possible in a financially sustainable way, should choose high status employees.
Furthermore, an important control variable in our study is salary. After all, higher salary levels
should, at least in theory, boost productivity (Akerlof & Yellen, 1986). Thus, it can be argued that
the significantly positive productivity effects found for the HH and HL matches are driven by
higher salaries. Including the salary control variable enabled us to better interpret the results on
the matching variables. Yet, despite the theoretical assumption, the salary variable was
insignificant in our full sample as well as in the regional regression analysis. Paying higher salaries
in MFIs seems not to enhance productivity, all else equal.
18
From the productivity/economics perspective, our finding on the second research question that
high status employees are more productive than low status employees irrespective of client type,
may seem close to a tautology- ‘more productive employees are more productive’. However,
referring to the homophily effect and similarity attractiveness paradigm, it is not obvious that high
status employees would be the most productive irrespective of the clients served. Indeed, as we
argue in the theoretical framework, if similarity attractiveness effects are strong, low status
employees may be more productive in serving low status clients. An interpretation of our findings
is that the productivity effect dwarfs the similarity effect. Nevertheless, based on results of the
second research question, one may question why MFIs then match on the LL category? After all,
we find that productivity would have been high if the MFI had selected high-status staff. This
suggests that there are other benefits (not captured in the labour productivity measure) to having a
low status similarity match.
Our finding that firms serving BOP markets, in our case MFIs, will benefit from hiring people
from a high socio-economic status regardless of their clients’ status is somewhat worrying. It
suggests that these firms are less likely to recruit from the poorer communities. Although offering
employment is secondary to microfinance’s social goal, ignoring it, may reinforce labour market
inequalities in these BOP markets. To cope with this, such firms could benefit from adopting status
enhancing mechanisms for potential local staff. This could be done through appropriate training,
salary, and other status-enhancing mechanisms. Moreover, the reader should keep in mind that the
microlending business involves complex operations including calculation of risk. In less
competence-demanding BOP markets, the importance of hiring high-status employees may be
lower.
Our results from robustness checks allude to potential regional differences on the impact of the
employee-client socio-economic categories. In Latin America & the Caribbean and South East
Asia & Pacific regions, our results are the same as in the full sample with the HL category (i.e.,
employees of higher status than clients) being the best performing in terms of productivity. Several
reasons could explain this. Firstly, these regions tend to be characterised by a high-power distance
culture (Sweetman, 2012; Gomez & Sanchez, 2005). This implies that vast inequalities among
individuals are considered normal in the society such that lower status individuals look up to higher
status individuals. Secondly, the regions tend to be characterized by high income disparities
19
relative to others (Wu & Chang, 2019; Amarante, Galván & Mancero, 2016). Such income
disparities imply differences in accessibility to health services and education opportunities, hence
reinforcing socio-economic differentials (Kraus et al., 2013). Therefore, these reasons may explain
why high-status staff are likely to perform better regardless of the client’s socio-economic status.
On the contrary, findings from Sub-Saharan Africa, Europe & Central Asia and the Middle East
& North Africa suggest that employee-client socio-economic matching does not matter for
productivity. These regions tend to have lower power distance in comparison to Latin America &
the Carribean and South-East Asia & Pacific. This suggests that there is less inequality among
individuals and may therefore explain why socio-economic status similarities or the lack thereof
do not seem to matter for performance. Additionally, differences in levels of institutional
development may offer some explanation, for instance, there is less dependence on relationship-
based interactions in relatively strong institutional environments (Boehe & Cruz, 2013). This
suggests that individual outcomes may be less influenced by status similarities or the lack thereof
in more developed regions like Europe and Central Asia. Furthermore, the high degree of diversity
between and within countries in some regions like Sub-Saharan Africa and Middle East & North
Africa (Green, 2013; Alesina et al., 2003) suggests that other status related aspects may be at play
such as tribe, clan or religion.
Finally, aspects of commercialization of microfinance may be informative to the overall
discussion. Studies show that the commercialization of microfinance firms has led to operational
changes such as shifts in lending methods from group to individual lending (de Quidt, Fetzer &
Ghatak, 2018) and increase in average loan sizes disbursed to clients (D’Espallier et al., 2017).
This begs the question as to whether organizations change their employees when they
commercialize. Also, do they then match employees to meet their new target clients? If so, this
may reinforce the need for high status employees. Although a few scholars have suggested that
commercialization may lead to human resource changes (Ledgerwood & White, 2006), the extent
to which commercialization is related to employment practices and who to hire is intriguing. In
our study, we do not distinguish between employee-client categories before and after
commercialization. Extending our analysis in this direction is an interesting topic for future
research.
20
7. Conclusion
The objectives of this study were twofold. One was to determine whether there are employee-client
matches based on socio-economic status in BOP firms. The other was to determine whether similar
socio-economic matching yields favourable employee productivity.
The results show that microfinance institutions tend to match their employees and clients based on
socio-economic status. 58 per cent of the employee- client categorizations were based on similarity
whereas 42 per cent were based on dissimilarity. High socio-economic status employees were
found more suitable for employee productivity both when targeting high and low socio-economic
status clients. The least favourable category was employees of low socio-economic status serving
clients of high socio-economic status. Nevertheless, additional results suggest the impact of the
socio-economic matching categories on productivity is dependent on the region.
One of microfinance’s specificities includes the important role that employees play as mediators
in achieving firm objectives (Siwale & Ritchie, 2012). Yet, the underlying factors within loan
officer- client categorizations remain scarcely explored. Our study attempts to contribute to this
by exploring the compatibility of loan officers and clients based on their socio-economic status
and how this impacts productivity. This study suggests that focus should not merely be on the
establishment of the relationship but also, on understanding how social aspects between loan
officers and clients that influence the development of such relationships hence, performance.
Although the BOP market has garnered a lot of attention, there is still scarcity of human resource
related research. Our study attempts to fill this gap in mainly two ways. First, it contributes to
matching-related studies in general human resource literature by exploring the concept in the BOP
context. Secondly, our study may have implications for the employment strategy of microfinance
institutions. Particularly, since high-status employees appear to be in high demand, managers have
to think strategically about the matching of clients and employees as its impact may vary by
geographical location. Due to cultural and institutional differences in the regions, it is necessary
for each MFI to carefully consider the that type of employees that they need in order to be effective
in their context. In some contexts, they may perform better by hiring higher status staff whereas in
21
other contexts the employees’ status may not matter. This may be relevant especially for firms
participating in local engagement through hiring locals among their staff.
Moreover, microfinance institutions represent just one example of social enterprise thus, the
employment strategy among different social enterprises may differ. For example, some social
enterprises tend to have both volunteers and paid employees among their staff (Doherty et al.,
2014). A question may therefore be raised as to whether a selection criterion for each employee
type should be in place? Also, in other BOP enterprises, whether it matters what type of clients the
firm is targeting, i.e., paying clients or beneficiaries. Our study therefore opens the avenue for
future research that seeks to explain such human resource related nuances in social enterprises.
Moreover, it also raises questions as to whether upcoming issues in social entrepreneurship such
as commercialization may influence the firms’ employment strategy
Future studies can also consider specific individual-level socio-economic information on the
employees and clients to establish whether their findings differ from those of our study which
employs firm level data. Other related aspects of socio-economic status for example, the education
level, tribe, clan of the employees and clients, can also be investigated in future BOP employee-
client matching-related studies. Likewise, investigating other mechanisms inherent in employee-
client interactions beyond socio-economic status and across different regions based on cultural
dimensions like power distance could be informative.
Notes
1. Microfinance practitioners inform that MFIs for instance those in Latin America often
disburse large average loan sizes per client of $5000 or more, making clients likely to be
wealthier than employees.
2. Although rated data is generally considered to be authentic, it tends to lack small saccos as
well as bigger microfinance banks. It is also over-represented in the Latin American region.
Ultimately, no dataset is a perfect representation of the population and the rating dataset is no
exception to this.
22
3. For each category, the letter in the first position represents the employee’s socio-economic
status and the letter in second position represents the client’s status. For example, in the HL
category the employee is of high socio-economic status and the client is of low socio-
economic status.
4. We also ran similar regressions on other MFI performance variables such as profit, financial
revenue and portfolio yield and did not find any significant effects (results are available upon
request).Therefore, we consider productivity the most relevant in this study since matching
is most likely to influence it and to a lesser extent profit, financial revenue and portfolio yield
which are more dependent on other market factors.
Funding details: This research did not receive any specific grant from funding agencies in the
public, commercial, or not-for-profit sectors.
Disclosure Statement: No potential conflict of interest was reported by the authors.
References
Abate, G. T., Borzaga, C., & Getnet, K. (2014). Cost‐efficiency and outreach of microfinance institutions:
Trade‐offs and the role of ownership. Journal of International Development, 26(6), 923–932.
Adler, N. E., & Snibbe, A. C. (2003). The role of psychosocial processes in explaining the gradient
between socioeconomic status and health. Current Directions in Psychological Science, 12(4),
119–123.
Ahmad, M. M. (2002). NGO Fieldworkers in Bangladesh. Aldershot, UK: Ashgate Publishing Company.
Akerlof, G. A., & Yellen, J. L. (eds.). (1986). Efficiency Wage Models of the Labor Market. Cambridge
University Press.
Alesina, A., Devleeschauwer, A., Easterly, W., Kurlat, S., & Wacziarg, R. (2003).
Fractionalization. Journal of Economic growth, 8(2), 155-194.
Amarante, V., Galván, M., & Mancero, X. (2016). Inequality in Latin America: a global
measurement. Cepal Review, 2016(118), 25-44.
Arellano, M. (1987). Computing robust standard errors for within-groups estimators. Oxford Bulletin of
Economics and Statistics, 49, 431–434.
23
Armendáriz, B., & Szafarz, A. (2011). On mission drift in microfinance institutions. In The Handbook of
Microfinance (pp. 341–366). Singapore: World Scientific.
Avery, D. R., McKay, P. F., Tonidandel, S., Volpone, S. D., & Morris, M. A. (2012). Is there method to
the madness? Examining how racioethnic matching influences retail store productivity. Personnel
Psychology, 65(1), 167–199.
Banthia, A., Greene, J., Kawas, C., Lynch, E., & Slama, J. (2011). Solutions for financial inclusion:
Serving rural women. Women’s World Banking Focus Note. New York: Women’s World
Banking.
Battilana, J., & Dorado, S. (2010). Building sustainable hybrid organizations: The case of commercial
microfinance organizations. Academy of Management Journal, 53(6), 1419–1440.
Battilana, J., Sengul, M., Pache, A. C., & Model, J. (2015). Harnessing productive tensions in hybrid
organizations: The case of work integration social enterprises. Academy of Management
journal, 58(6), 1658-1685.
Beck, T. H. L., Behr, P., & Madestam, A. (2011). Sex and credit: Is there a gender bias in microfinance?
Discussion paper 2011-101, Tilburg University.
Beisland, L. A., D’Espallier, B., & Mersland, R. (2019). The commercialization of the microfinance
industry: Is there a ‘personal mission drift’among credit officers?. Journal of Business
Ethics, 158(1), 119-134.
Bell, R., Harper, A., & Mandivenga, D. (2002). Can commercial banks do microfinance? Lessons from
the Commercial Bank of Zimbabwe and the Co-operative Bank of Kenya. Small enterprise
development, 13(4), 35-46.
Bocken, N. M., Fil, A., & Prabhu, J. (2016). Scaling up social businesses in developing markets. Journal
of Cleaner Production, 139, 295-308.
Boehe, D. M., & Cruz, L. B. (2013). Gender and microfinance performance: Why does the institutional
context matter?. World Development, 47, 121-135.
Bruns, V., Holland, D. V., Shepherd, D. A., & Wiklund, J. (2008). The role of human capital in loan
officers’ decision policies. Entrepreneurship Theory and Practice, 32(3), 485–506.
Bocken, N. M., Fil, A., & Prabhu, J. (2016). Scaling up social businesses in developing markets. Journal
of Cleaner Production, 139, 295-308.
Byrne, D. (1969). Attitudes and attraction. In Advances in Experimental Social Psychology (Vol. 4, pp.
35–89). Academic Press.
Byrne, D. (1971). The Attraction Paradigm. New York: Academic Press.
Byrne, D., Clore, G. L. Jr., & Worchel, P. (1966). Effect of economic similarity-dissimilarity on
interpersonal attraction. Journal of Personality and Social Psychology, 4(2), 220–224.
24
Chhibber, A., & Nayyar, G. (2008). Pro‐poor growth: explaining the cross‐country variation in the growth
elasticity of poverty. International Journal of Development Issues.
Copestake, J. (2007). Mainstreaming microfinance: Social performance management or mission
drift?. World development, 35(10), 1721-1738.
Cox, T. H., & Blake, S. (1991). Managing cultural diversity: Implications for organizational
competitiveness. Academy of Management Perspectives, 5(3), 45–56.
Cull, R., Demirguz-Kunt, A. & Morduch, J. (2007). Financial performance and outreach: A global analysis
of leading microbanks. The Economic Journal, 117, F107–F133.
Cull, R, Demirgüç-Kunt, A. & Morduch, J. (2009). Microfinance meets the market. Journal of Economic
Perspectives 23(1), 167–92.
D’Espallier, B., Goedecke, J., Hudon, M., & Mersland, R. (2017). From NGOs to banks: Does institutional
transformation alter the business model of microfinance institutions?. World Development, 89, 19-
33.
De Quidt, J., Fetzer, T., & Ghatak, M. (2018). Commercialization and the decline of joint liability
microcredit. Journal of Development Economics, 134, 209-225.
Doherty, B., Haugh, H., & Lyon, F. (2014). Social enterprises as hybrid organizations: A review and
research agenda. International journal of management reviews, 16(4), 417-436.
Dwyer, S., Orlando, R., & Shepherd, C. D. (1998). An exploratory study of gender and age matching in
the salesperson-prospective customer dyad: Testing similarity-performance predictions. Journal
of Personal Selling & Sales Management, 18(4), 55–69.
Fisman, R., Paravisini, D., & Vig, V. (2017). Cultural proximity and loan outcomes. American Economic
Review, 107(2), 457–492.
Garmaise, M. J., & Natividad, G. (2010). Information, the cost of credit, and operational efficiency: An
empirical study of microfinance. Review of Financial Studies, 23(6), 2560–2590.
Galak, J., Small, D., & Stephen, A. T. (2011). Microfinance decision making: A field study of prosocial
lending. Journal of Marketing Research, 48, S130–S137.
Gomez, C., & Sanchez, J. I. (2005). HR's strategic role within MNCs: helping build social capital in Latin
America. The international journal of human resource management, 16(12), 2189-2200.
Gonzalez, J. A. (2013). Matchmaking: Community and business unit racial/ethnic diversity and business
unit performance. International Journal of Human Resource Management, 24(21), 4063–4081.
Goyal, S., Sergi, B. S., & Jaiswal, M. P. (2016). Understanding the challenges and strategic actions of
social entrepreneurship at base of the pyramid. Management Decision.
Gray, B., & Kish-Gephart, J. J. (2013). Encountering social class differences at work: How “class work”
perpetuates inequality. Academy of Management Review, 38(4), 670-699.
25
Green, E. (2013). Explaining African ethnic diversity. International Political Science Review, 34(3), 235-
253.
Gutiérrez-Nieto, B., & Serrano-Cinca, C. (2007). Factors explaining the rating of microfinance
institutions. Nonprofit and Voluntary Sector Quarterly, 36(3), 439–464.
Hammond, A.L., Kramer, W.J., Katz, R.S., Tran, J.T. (2007). The next 4 billion: market size and business
strategy at the base of the pyramid. World Bank Group. Available at
http://documents.worldbank.org/curated/en/779321468175731439/The-next-4-billion-market-
size-and-business-strategy-at-the-base-of-the-pyramid Accessed on 14th August 2020.
Hart, S. L., & London, T. (2005). Developing native capability. Stanford Social Innovation Review, 3(2),
28–33.
Hartarska, V., Shen, X., & Mersland, R. (2013). Scale economies and input price elasticities in
microfinance institutions. Journal of Banking & Finance, 37(1), 118–131.
Hermes, N., Lensink, R., & Meesters, A. (2011). Outreach and efficiency of microfinance
institutions. World development, 39(6), 938-948.
Hofstede, G. (2001). Culture's consequences: Comparing values, behaviors, institutions and
organizations across nations. Sage publications.
Hofstede, G. H., Hofstede, G. J., & Minkov, M. (2005). Cultures and organizations: Software of the
mind (Vol. 2). New York: Mcgraw-hill.
Hudon, M., & Traca, D. (2011). On the efficiency effects of subsidies in microfinance: An empirical
inquiry. World Development, 39(6), 966–973.
Jacobs, J., & Franceys, R. (2008). Better practice in supplying water to the poor in global public private
partnerships. Proceedings of the Institution of Civil Engineers – Municipal Engineer,161(4), 247–
254.
Joshi, A., & Roh, H. (2009). The role of context in work team diversity research: A meta-analytic review.
Academy of Management Journal, 52(3), 599–627.
Jones, E., Moore, J. N., Stanaland, A. J., & Wyatt, R. A. (1998). Salesperson race and gender and the
access and legitimacy paradigm: Does difference make a difference? Journal of Personal Selling
& Sales Management, 18(4), 71–88.
Kennedy, N.S. (2012, June 19). The challenges and opportunities of hiring locals. Devex- International
Development News. Retrieved from https://www.devex.com/news/the-challenges-and-
opportunities-of-hiring-locals-78472
Kochan, T., Bezrukova, K., Ely, R., Jackson, S., Joshi, A., Jehn, K., Leonard, J., Levine, D., & Thomas,
D. (2003). The effects of diversity on business performance: Report of the diversity research
network. Human Resource Management, 42(1), 3–21.
26
Kraus, M. W., Tan, J. J., & Tannenbaum, M. B. (2013). The social ladder: A rank-based perspective on
social class. Psychological Inquiry, 24(2), 81-96.
Labie, M., Méon, P. G., Mersland, R., & Szafarz, A. (2015). Discrimination by microcredit officers:
Theory and evidence on disability in Uganda. Quarterly Review of Economics and Finance, 58,
44–55.
Ledgerwood, J., & White, V. (2006). Transforming microfinance institutions: providing full financial
services to the poor. The World Bank.
Lensink, R., Mersland, R., Vu, N. T. H., & Zamore, S. (2018). Do microfinance institutions benefit from
integrating financial and nonfinancial services? Applied Economics, 50(21), 2386–2401.
Loewenstein, G., & Small, D. A. (2007). The scarecrow and the tin man: The vicissitudes of human
sympathy and caring. Review of General Psychology, 11(2), 112.
McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social
networks. Annual Review of Sociology, 27(1), 415–444.
Mersland, R., & Strøm, R. Ø. (2010). Microfinance mission drift? World Development, 38(1), 28–36.
MicroRate. (2014). Technical Guide: Performance and social Indicators for microfinance Institutions.
Retrieved 10.04.19 from http://www.microrate.com.
Morduch, J. (1999). The microfinance promise. Journal of economic literature, 37(4), 1569-1614.
Morrison, A. M. (1992). The New Leaders: Guidelines on Leadership Diversity in America. Jossey-Bass
Management Series. San Francisco: Jossey-Bass Publishers.
Oakes, J. M., & Rossi, P. H. (2003). The measurement of SES in health research: Current practice and
steps toward a new approach. Social Science & Medicine, 56(4), 769–784.
Parekh, N., & Ashta, A. (2018). An institutional logics perspective to evolution of Indian microcredit
business models. Strategic Change, 27(4), 313-327.
Périlleux, A., Hudon, M., & Bloy, E. (2012). Surplus distribution in microfinance: Differences among
cooperative, nonprofit, and shareholder forms of ownership. Nonprofit and Voluntary Sector
Quarterly, 41(3), 386-404.
Prahalad, C. K. (2006). The Fortune at the Bottom of the Pyramid. Upper Saddle River, NJ: Pearson
Education.
Prahalad, C. K., & Hart, S. L. (2002). The fortune at the bottom of the pyramid. Strategy & Business, 26.
Rai, A., Maruping, L. M., & Venkatesh, V. (2009). Offshore information systems project success: The
role of social embeddedness and cultural characteristics. MIS Quarterly, 33(3), 617–641.
Randøy, T., Strøm, R. Ø., & Mersland, R. (2015). The impact of entrepreneur–CEOs in microfinance
institutions: A global survey. Entrepreneurship Theory and Practice, 39(4), 927-953.
27
SadreGhazi, S., & Duysters, G. (2009). Serving low-income markets: rethinking multinational
corporations’ strategies. In Dolfsma, W., Duysters, & Costa, I. (eds), Multinationals and Emerging
Economies, pp. 166–182.
Schreiner, M. (2002). Aspects of outreach: A framework for discussion of the social benefits of
microfinance. Journal of international development, 14(5), 591-603.
Shah, A. K., Mullainathan, S., & Shafir, E. (2012). Some consequences of having too
little. Science, 338(6107), 682-685.
Siwale, J. (2016). Microfinance and loan officers’ work experiences: Perspectives from Zambia. Journal
of Development Studies, 52(9), 1289–1305.
Siwale, J. N., & Ritchie, J. (2012). Disclosing the loan officer’s role in microfinance development.
International Small Business Journal, 30(4), 432–450.
Smith, J. B. (1998). Buyer-seller relationships: Similarity, relationship management, and quality.
Psychology & Marketing, 15(1), 3–21.
Smith, W. K., Gonin, M., & Besharov, M. L. (2013). Managing social-business tensions: A review and
research agenda for social enterprise. Business Ethics Quarterly, 23(3), 407-442.
Stephens, N. M., Fryberg, S. A., & Markus, H. R. (2011). When choice does not equal freedom: A
sociocultural analysis of agency in working-class American contexts. Social Psychological and
Personality Science, 2(1), 33-41.
Sweetman, K. (2012, April 10). In Asia, Power Gets in the Way. Harvard Business Review.
https://hbr.org/2012/04/in-asia-power-gets-in-the-way
Schuster, T., & Holtbrügge, D. (2012). Market entry of multinational companies in markets at the bottom
of the pyramid: A learning perspective. International Business Review, 21(5), 817-830.
Van Den Berg, M., Lensink, R., & Servin, R. (2015). Loan officers’ gender and microfinance repayment
rates. Journal of Development Studies, 51(9), 1241–1254.
Willems, S., Stéphanie de Maesschalck, S., Deveugele, M., Derese, A., & Maeseneer, J. (2005). Socio-
economic status of the patient and doctor–patient communication: Does it make a difference?
Patient Education and Counseling 56 (2), 139–146.
Williams, J., & Nguyen, N. (2005). Financial liberalisation, crisis, and restructuring: A comparative study
of bank performance and bank governance in South East Asia. Journal of Banking & Finance,
29(8-9), 2119–2154.
Wry, T., & Zhao, E. Y. (2018). Taking trade-offs seriously: Examining the contextually contingent
relationship between social outreach intensity and financial sustainability in global microfinance.
Organization Science, 29(3), 1–23.
Wu, W. C., & Chang, Y. T. (2019). Income inequality, distributive unfairness, and support for democracy:
evidence from East Asia and Latin America. Democratization, 26(8), 1475-1492.
28
Zamore, S. (2018). Should microfinance institutions diversify or focus? A global analysis. Research in
International Business and Finance, 46, 105–119.
Tables
Table 1. Summary of BOP Firm Employee-Client Matches
High Socio-economic Status
Clients
Low Socio-economic Status Clients
High Socio-economic
Status Employees
High-Status Match (HH)
(1)
Employees Serving Downwards (HL)
(2)
Low Socio-economic
Status Employees
Employees Serving Upwards (LH)
(3)
Low-Status Match (LL)
(4)
Notes: For the abbreviations HH, LH, HL and LL, the letter in the first position represents the employees’ status, whereas the
letter in the second position represents the clients’ status
Table 2. Descriptive Statistics
Variables
Description
Mean
Std.
Dev
Min
Max
Matching Variables: Equation (1)
HH
High-status match
0.30
0.46
0
1
LL
Low-status match
0.28
0.45
0
1
HL
Employees serving
downwards
0.21
0.41
0
1
LH
Employees serving upwards
0.21
0.40
0
1
Dependent Variable: Equation (2)
Employee
Productivity
Number of credit clients per
MFI employee
118.24
74.91
9
396
Control Variables: Equation (2)
Salary
Average salary: Total
employee costs divided by
number of MFI employees
1.13
1.46
0.04
8.92
29
Loansize
Average loan size: Total
loan portfolio divided by
number of credit clients
0.17
0.37
0.00
2.57
Risk
Total risk of the MFI
(par30+writeoff)
0.08
0.10
0
0.60
MFI-size
Natural Logarithm of Total
assets from the balance
sheet
2309.77
5631.14
7.84
37640.87
MFI-age
Years since establishment
of the MFI
12.61
8.94
1
45
Urban
market
1 = MFI has some
operations in urban market,
0 = otherwise
0.88
0.31
0
1
Shareholder
1= shareholder owned, 0 =
otherwise
0.37
0.48
0
1
Creditmethod
1= individual lending as
main method, 0= otherwise
0.57
0.50
0
1
Note: All monetary variables are deflated by GDP per capita, PPP $
Table 3. Socio-economic Matching Frequencies
HH
LL
HL
LH
µ
0.30***
0.28***
0.21***
0.21***
(0.01)
(0.01)
(0.01)
(0.01)
H0: µ=0.25
0.000001
0.002
0.00003
0.000002
Observations
1923
1923
1923
1923
Notes: HH represents high-status match; LL, is the low-status match; HL, is employees serving downwards; and LH,
is employees serving upwards. The figures of µ show probability values of the test µ = 0.25. Standard errors are in
parentheses. *, ** and *** denote statistical significance at the 10%, 5% and 1% level respectively.
30
Table 4. Performance Regressions
Employee Productivity
(1)
(2)
HH
10.35***
11.16***
(2.94)
(2.80)
LL
11.60***
7.263**
(3.73)
(3.31)
HL
24.27***
21.12***
(4.30)
(4.16)
Salary
-3.47
(4.26)
Loansize
-64.73***
(10.45)
MFI Size
11.35***
(3.15)
MFI Age
-5.649
(3.72)
Urban market
-14.54*
(8.47)
Risk
-59.95***
(15.87)
Shareholder-
owned
-2.296
(15.87)
Creditmethod
-4.39
(5.20)
Country specific
time trend
Yes
Yes
R2
0.18
0.25
Observations
1923
1923
MFIs
474
474
H0: LL=HL
0.0008
0.0002
H0: HH=HL
0.0006
0.009
Notes: Standard errors in parentheses. *, ** and *** denote statistical significance at the 10%, 5% and 1% level
respectively.H0: LL=HL indicates values of the probability that the coefficient of LL is not different from HL. H0:
HH=HL indicates values of the probability that the coefficient of HH is not different from HL.
31
Table 5: Regional Performance Regressions
(1)
(2)
(3)
(4)
(5)
VARIABLES
LAC
SSA
ECA
MENA
SEAP
HH
11.94***
3.703
-0.379
10.07
35.69**
(4.132)
(7.333)
(4.723)
(15.95)
(17.09)
LL
10.41**
0.214
-8.462*
-11.41
47.61*
(4.545)
(12.46)
(4.839)
(14.15)
(25.27)
HWLL
21.70***
-9.212
10.60*
23.72
57.24***
(5.613)
(14.87)
(5.517)
(15.05)
(16.64)
Salary
-4.755
1.387
5.389
-7.582
10.05*
(5.689)
(5.785)
(3.795)
(15.94)
(5.106)
loansize
-122.7***
-22.72**
-38.61***
-542.7**
-68.55
(31.17)
(9.421)
(12.07)
(245.5)
(48.31)
MFI Size
17.10***
23.54***
3.417
8.660
-0.561
(4.820)
(8.656)
(3.011)
(17.64)
(7.838)
MFI Age
-6.825**
-20.50
-8.495
5.112
-3.439
(2.968)
(22.13)
(11.35)
(4.533)
(3.266)
Urban market
-5.938
-28.02**
-1.529
-43.01
(9.183)
(12.45)
(9.527)
(47.04)
Risk
-28.58
-33.04
-41.87
-124.7***
-308.0
(23.49)
(22.25)
(38.67)
(34.69)
(224.6)
Shareholder-
owned
11.19
-13.11
2.381
-33.95
-59.16**
(24.81)
(16.25)
(14.45)
(22.89)
(21.44)
Creditmethod
1.839
-35.70**
-3.641
3.464
(5.155)
(15.29)
(10.30)
(9.106)
Constant
108.2***
151.8
44.12*
42.71
253.6***
(35.36)
(111.6)
(24.71)
(62.20)
(47.72)
Year dummies
Yes
Yes
Yes
Yes
Yes
R2
0.157
0.308
0.374
0.700
0.535
Observations
1,053
378
305
94
93
MFIs
225
119
82
22
26
Notes: Standard errors in parentheses. *, ** and *** denote statistical significance at the 10%, 5% and 1% level
respectively. LAC represents Latin America & the Caribbean, SSA represents Sub-Saharan Africa. ECA represents
Europe & Central Asia, MENA represents the Middle East & North Africa and SEAP represents South East Asia &
Pacific.
32
Appendix 1: Distribution of MFIs by Country and Region
SSA
LAC
ECA
#
Country
No. of
MFIs
#
Country
No. of
MFIs
#
Country
No. of
MFIs
1
Angola
1
37
Argentina
2
69
Albania
3
2
Benin
8
38
Bolivia
18
70
Armenia
6
3
Burkina Faso
9
39
Brazil
25
71
Azerbaijan
9
4
Burundi
6
40
Chile
2
72
Bosnia and
Herzegovina
12
5
Cameroon
6
41
Colombia
14
73
Bulgaria
3
6
Chad
2
42
Costa Rica
3
74
Croatia
1
7
Comoros
1
43
Dominican
Republic
8
75
Georgia
8
8
Congo
1
44
Ecuador
24
76
Italy
3
9
Dem Rep
Congo
2
45
El Salvador
8
77
Kazakhstan
8
10
Ethiopia
10
46
Guatemala
10
78
Kyrgyzstan
9
11
Gambia
1
47
Haiti
3
79
Moldova
2
12
Ghana
6
48
Honduras
18
80
Montenegro
2
13
Guinea
3
49
Jamaica
1
81
North Macedonia
1
14
Kenya
18
50
Mexico
32
82
Romania
7
15
Madagascar
4
51
Nicaragua
19
83
Russia
17
16
Malawi
3
52
Paraguay
2
84
Serbia
2
17
Mali
11
53
Peru
47
85
Tajikistan
11
18
Mozambique
1
54
Trinidad &
Tobago
1
86
Turkey
1
19
Niger
9
SEAP
87
United Kingdom
1
20
Nigeria
7
55
Afghanistan
2
Total MFIs
650
21
Rwanda
13
56
Bangladesh
2
22
Senegal
12
57
Cambodia
14
23
Sierra Leone
2
58
China
5
24
South Africa
4
59
India
32
25
South Sudan
1
60
Indonesia
4
26
Tanzania
9
61
Mongolia
4
27
Togo
5
62
Nepal
5
28
Uganda
25
63
Pakistan
2
29
Zambia
3
64
Philippines
22
MENA
65
Samoa
1
30
Egypt
6
66
Sri Lanka
2
31
Jordan
3
67
Timor-Leste
1
32
Lebanon
2
68
Viet Nam
4
33
Morocco
8
34
Palestine
3
35
Tunisia
1
36
Yemen
1
Notes: LAC represents Latin America & the Caribbean, SSA represents Sub-Saharan Africa. ECA represents Europe
and Central Africa, MENA represents the Middle East & North Africa and SEAP represents South East Asia & Pacific.
