Whose gain is it anyway? Structurational perspectives on deploying ICTs for development in India's...

Whose Gain Is It Anyway? StructurationalPerspectives on Deploying ICTs for Developmentin India’s Microfinance Sector

Rahul De’Indian Institute of Management Bangalore, Bangalore – 560076, India.E-mail: [email protected]

Aishwarya Lakshmi RatanMicrosoft Research India, “Scientia,” Sadashivnagar, Bangalore – 560080, India.E-mail: [email protected]

ABSTRACT

The idea of information and communication technology (ICT) being a “hammer” that can be applied toa wide variety of “nails” across different geographic locations, sectors, organizations, and contexts toimprove efficiency and/or have a beneficial social impact has come under severe criticism, particularlyin the realm of implementing socioeconomic development programs. Structuration theory remainsone of the key metatheories that deconstruct the complexity of technology introductions in thecontext of organizational and behavioral change. In this study, we use a structurational lens toexamine two pilot ICT implementations in the Indian microfinance sector, specifically exploring theinteractions between the ICT intervention, the organizations and people implementing the change,and the structural and institutional context within which these projects were rolled out. We showcasehow an “ICT for development” intervention is inherently a political process, involving choices arounddefining efficiency and targeting particular social welfare improvements, with varying repercussionsfor the involved microfinance institution and client. Where the client’s context, constraints, andwelfare are placed at the heart of the “efficiency” discussion during the technology’s design andimplementation, the development impact is seen to be far greater and more sustained. C© 2009WileyPeriodicals, Inc.

Keywords: Information and communication technologies; development; microfinance; self-helpgroups; efficiency; organizational change; structuration theory; India

1. INTRODUCTION

There appears to be widespread agreement among both researchers and policy makersthat information and communication technologies (ICTs) are necessary for developingcountries, whether for creating businesses around ICT products (telecenters, rural computerkiosks, urban internet cafes, community radio stations), or for using ICTs to supportvarious commercial, developmental, and governance activities. There is also widespreadacknowledgment that when it comes to either of these aspects of ICT deployment and use

Yola Georgiadou is the accepting Editor for this article.

Information Technology for Development, Vol. 15 (4) 259–282 (2009) C© 2009 Wiley Periodicals, Inc.Published online 10 September 2009 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/itdj.20129

259

260 DE’ AND RATAN

toward achieving socioeconomic development goals, the benefits from interventions arefar from automatically realized, and the number and range of failures is notoriously largecompared to the successes. It is imperative, thus, to understand the nature and complexityof information systems (IS) implementations in socioeconomic development efforts, wherebudgets are constrained and the potential impact involves not only efficiency gains but alsodevelopment outcomes.

This article examines the use of ICTs in the context of the microfinance sector with a viewto examine, in detail, the particularities of design, deployment, and organizational practicesthat are associated with positive outcomes. Microfinance institutions (MFIs) and facilitatorshave gained prominence over the past two decades as drivers of key development initiativesin the financial sector, given that they usually have a wide reach in rural areas that aretraditionally underserved by the formal financial sector (Armendariz & Morduch, 2005).The evidence so far points to MFIs having a net positive impact on particular sections ofpoor populations under particular enabling conditions (Emran, Morshed, & Stiglitz, 2007;Haqqani, 2003; Khandker, 2005).

However, few MFIs have achieved full financial self-sufficiency, partly due to their highoperating costs. One of the key challenges they face is that of reducing client transactionscosts related to borrowing, lending, accounting, reporting, control, etc. (Bhatt & Tang, 1998;Kamel, 2005). Owing to the large volume and high frequency of geographically dispersedtransactions, but low values associated with each transaction, client transaction costs haveto be controlled in order to ensure financial viability for both lenders and borrowers (Ivatury,2006; Parikh, 2006). Recourse to ICT solutions has, therefore, been an attractive propositionfor several large MFIs (Ananth et al., 2004; Mathison, 2005; Regy & Mahajan, 2006).

We examine ICT introductions in two MFIs located in India. Both institutions are ofcomparable size and nature, and introduced ICTs to improve the efficiency of tasks inthe realm of client transaction management. However, one project endured and has scaledacross the MFI’s branches nationwide, whereas the other was withdrawn after the pilotimplementation. These different results raise many issues, including those of context,organizational setting, technology design, implementation process, and participation, inaffecting project outcomes.

Specifically, we find that an “ICT for development” intervention is inherently a politicalprocess, involving choices around defining efficiency and targeting particular social welfareimprovements, with varying repercussions for the involved MFI and client. Where theclient’s context, constraints, and welfare are placed at the heart of the “efficiency” discussionduring the technology’s design and implementation, the development impact is seen to befar greater and more sustained. This is a major departure from the dominant rhetoric in theICT for development domain, in which technologies themselves are taken to be apolitical.Following Stone (2002), this is also a departure from the neoclassical view that there is asingular and objective notion of “efficiency” and that ICT-enabled processes are inherentlyefficiency enhancing and consequently welfare improving for all involved stakeholders.

We proceed as follows in this article: in section 2, we first present an overview ofrelated work in deploying ICTs for development, and then present structuration as a broadtheoretical framework through which we examine our cases. We then present an overview ofmicrofinance operations in section 3. This framework allows us to draw up a set of researchquestions that guides our analysis in the ICT for microfinance domain in section 4.1, whichis followed by a discussion of our research methodology in section 4.2. We then describeour two cases in sections 5 and 6. In section 7, we explore our initial research questionsthrough a fine-grained analysis and discussion of the cases. We conclude with thoughts on

Information Technology for Development DOI: 10.1002/itdj

WHOSE GAIN IS IT ANYWAY? 261

the implications of our findings in section 8, and summarize the key contributions of thisresearch to the literature and future work in section 9.

2. THEORETICAL FRAMEWORK

There are a few broad themes under which much of the research in ICTs for developmentis conducted (Walsham, Robey, & Sahay, 2007; Walsham & Sahay, 2005). These themesinclude the role of ICTs in overall economic development; the use of ICTs for the upliftmentof marginalized groups and its impact; the design and implementation impact of particularICT artifacts; and issues related to the theory of technology for development research,such as methodological stance and level of analysis. This article focuses on the group ororganizational level of analysis and on the issues related to the introduction of an ICTartifact. Prior research related to our work can be grouped into three themes—ICT use inmicrofinance, ICT deployment in rural/nonurban areas, and the contextual conditions ofICT use. We provide a brief overview of the research in the next subsection. The followingsubsection briefly explains structuration theory that forms the theoretical basis for the article.

2.1 Prior Research on Deploying ICTs for Development

ICT use in microfinance has largely focused on the problem of efficiency enhancement,that is, reducing transactions costs for a large scale of operations (Kamel, 2005), thepotential for ICT introduction in streamlining the MFI’s operations (Frankiewicz, 2003;Parikh et al., 2006), the challenges facing distributed computing and channel requirements(Gurau, 2005; Kannabiran & Narayan, 2005), and the issues of scale and geographicallydistributed connectivity to sustain banking (Mauguis et al., 2005; Press, 2005).

ICT deployments in rural/nonurban areas have been quite widely studied, within applica-tion areas of health (Miscione, 2007), e-governance (Krishna & Walsham, 2005; Kuriyan,Ray, & Toyama, 2006), land information and management (De’, 2005; Puri, 2007), tele-centers or rural computer kiosks (Kiri & Menon, n.d.; Kumar, 2004; Parthasarathy, 2004;Rangaswamy, 2006), community radio (Bailur, 2007), and others. Common problems as-sociated with ICT deployments in rural environments are those of the lack of reliableelectronic connectivity, intermittent electrical power, maintenance difficulties, nonavail-ability of skilled personnel, low income and purchasing power, dearth of relevant ruralcontent/applications, low literacy levels for using technology, and inadequate social struc-tures for absorption of innovations. These constraints continue to persist in most rural areasof developing countries (Keniston, 2003). One of the challenges in ICT deployments is inovercoming these constraints to provide meaningful services.

The third aspect of the contextual use of ICT refers to the prevailing social, environmental,economic, and organizational conditions under which an ICT project is implemented. Thecontext is understood and defined by the level of analysis and the scope of the project. Ane-governance introduction in a large political space, for instance, has the context of thelocal political environment, the social and economic situation of the population affected bythe project, and the ideological mind-set of the implementers (Prakash & De’, 2007).

A deeper understanding of context derives from the theory of structuration that wasdeveloped to understand the implementation of IS in organizations (DeSanctis & Poole,2004; Jones & Karsten, 2008; Orlikowski & Robey, 1991). The structuration frameworkposits that ICTs both determine and are determinants of social conditions in any organizationin which they are adopted. The study of ICTs thus has to account for the role of human


262 DE’ AND RATAN

action in shaping the technology, along with the manner in which ICT changes and modifiessocial structures.

Where structuration emphasizes the social nature of technology adoption and use,Deborah Stone’s (2002) work on policy paradoxes adds a much-needed political dimensionto the examination of deploying ICTs for development. Although efficiency improvementsare considered uncontested motivations in pursuing technology deployments (particularlyin the microfinance sector), Stone (2002) questions any universal notion of efficiencyenhancement,

. . . there are any number of possible paths to the goal of [efficiency]. These paths, and their results,cannot be scored on any a priori metric of efficiency. Efficiency is always a contestable concept.Everyone supports the general idea of getting the most out of something, but to go beyond the valueslogans and apply the concept to a concrete policy choice requires making assumptions about whoand what counts as important. There are no correct answers to these questions to be found outside thepolitical process. The answers built into supposedly technical analyses of efficiency are nothing morethan political claims. By offering different assumptions, sides in a conflict can portray their preferredoutcomes as being most efficient. (pp. 65–66)

The framework we use invokes the theory of structuration and Stone’s work on policyparadoxes to describe the manner in which the complex interaction of ICTs, human actors,and institutions takes place.

2.2 Structuration Theory

Following Giddens’ (1984) original work on structuration theory, it has been used exten-sively in IS research. A recent study (Jones & Karsten, 2008) found more than 500 articlesappearing in major IS journals and conference proceedings that used or cited structurationtheory and, in particular, theorists of structuration such as Orlikowski, and DeSanctis andPoole. Structuration is one of the most widely used theoretical paradigms in IS literature(DeSanctis & Poole, 2004), and there are few situations in which IS is used that cannot bestudied within the paradigm of structuration.

Structuration relies on certain basic assumptions about humans and the situations theyinhabit: humans have agency and want to take actions in their particular situations and learnfrom them, humans have bounded knowledge and limited ability to know the consequencesof their actions, the context in which human interaction occurs is important and has tobe in the foreground, structural properties specify overall types of societies that enableunderstanding of the phenomena, and the study of power in interactions is important.

Structuration is based on a fundamental ontological assumption of duality: the phe-nomena of interest are studied from the perspective of humans and technology and theinteractions between them. Duality focuses on people and technology, how they are related,and how new forms of work and practice emerge from their interaction. This a priori as-sumption of duality enables the study of technology introductions within organizations and,more generally, the study of moments or instances at which technology is introduced. Dual-ity, arguably, has its limitations, where the choice of language or the discourse surroundingtechnology and people imposes limits on how certain “entanglements” (Orlikowski & Scott,2008) can be observed. In particular, the critics argue that duality precludes an understandingof the continuous and persistent use of technology and how it shapes work and organizations.

In the context of ICT introductions for specific purposes, in a developing country, itis useful to circumscribe the research within a specific context. Here, the structuration



assumptions are valid: there is a duality of technology being brought into the realm ofpeople/organizations; there is a moment of change, adaptation, or failure, all of whichare discontinuities that bear examination and throw light on the nature of the technologyintroduction; and there is an emergent nature of the study in which research questions areframed and reframed in the context of the study. Structuration theory does not detract fromthe object of the study, viz, why and under what conditions do ICT introductions succeed(because there is a long history of failures) and how can development be effected by suchtechnology introductions.

In the structurational model proposed by Orlikowski and Robey (1991), there are “. . . fourkey influences that operate continuously and simultaneously in the interaction betweentechnology and organizations” (p. 152). These four influences are ICTs as a product oroutcome of human action; ICTs as a means of human action, enabling action to take place;ICTs as constructed and used in a given social context; and ICT’s influence on the socialcontext within which it is constructed and used.

The first of these influences has two important implications: (1) that while technologyis designed or created, the social assumptions or beliefs of the creators will be reflectedin the resulting product; and (2) that technology is relevant and useful only when used.This latter aspect relates to the second influence, where the use of technology in daily workinfluences and changes the nature of the work, and this appropriation may in turn changethe work norms as they exist to a new form that was not predetermined by the technologydesign. The third influence refers to the fact that human actors are subject to the knowledge;resources; and structures of signification, domination, and legitimation of the organization(Orlikowski & Robey, 1991, p. 154). The fourth influence leads from the third that by usingICTs, human actors either substantiate or change organizational structures.

It is important to note that these influences act simultaneously and recursively(Orlikowski, 2000), impacting the consequences of each other and their own (from earlieractivations). The dynamic nature of structuration theory prevents its application directly,and the theory is considered as a metatheory meant to guide thinking and analysis of thephenomenon being examined.

We use structuration theory to draw up a framework of our own in section 4, whichis based on the particular context of microfinance (described in the next section), and onits basis, we pose a set of questions that examine the cross-influences of human actors,ICTs, and institutional properties on each other. The questions serve as a proxy for thesemetapropositions, and we examine them in light of the data obtained from our field study.

3. ICT USE IN MICROFINANCE

The term microfinance is used in this article to describe financial products and services thatare sold in small denominations to low-income clients (usually organized in groups), mostoften without physical collateral, whose good behavior (i.e., on-time repayment in the caseof microcredit) is ensured by relying on social networks and sanctions to prevent delin-quency and fraud. The provision of microfinance has been driven and promoted by a varietyof institutions ranging from nongovernmental organizations (NGOs), cooperatives, socialentrepreneurs, government agencies, and multilateral donors to large commercial banks.

The two microfinance organizations examined in this study are variants of the basicself-help group (SHG) model of microfinance that is practiced widely in India. A group ofusually between 10 and 20 women get together and make small (Rs. 5–20, ∼US$.12–.50),regular (usually weekly) contributions toward a common savings pool. This money is then


264 DE’ AND RATAN

continually re-lent to one or more of the group’s own members at a set interest rate (almostalways calculated on reducing balance). Central to the path to long-run sustainability is thelinkage of the SHG with a commercial bank or other third-party financial institution overa period of time. The idea is for the SHG to maintain its surplus finances in a joint savingsaccount at a commercial bank and eventually borrow larger amounts collectively fromformal banking institutions (at a low interest rate) for on-lending to individual members,thus removing the capital constraint faced by the SHG due to its own limited resources.Most SHGs earn on the spread between the external group rate offered by formal financialinstitutions (∼9% p.a.) and the internal individual interest rate they charge (12%–24% p.a.,depending on the maturity of the SHG). These earnings are periodically distributed tomembers as dividends (Ratan, 2007).

This model presents a win–win opportunity for all parties involved: it is one big loaninstead of multiple miniloans for the formal financial institution to service, it is a cheaperloan for the women borrowing than the external rates they individually face with privatefinanciers (>36% annual), and it has the added benefit of flexible principal repayment aslong as they make their regular interest payments on time. The women also benefit asco-owners of the high dividends earned on this pooled capital. This SHG–bank linkageprogram in India is one of the largest microfinance programs in the world, reaching outto more than 33 million member households (Punnathara, 2007). However, the quality ofthese groups is highly nebulous at present because there is no systematic and aggregatedrecord of their financial health and operations. Although supported by nodal governmentagencies, SHGs have been promoted by more than 3,000 different supporting institutionsacross the country (e.g., NGOs, state governments, religious institutions), each with its ownsystem of SHG promotion (Ratan, 2007).

There are two clear aspects to an SHG’s data management and processing functionsthat are separate but mutually dependent (i.e., management of internal individual financialtransactions and external group financial transactions). Each function has a different setof stakeholders and operational constraints, and often one or the other has assumed primeimportance in different microfinance settings. Given the nature of SHG operations, the bulkof cash transactions are internal because the lenders and borrowers are within the samegroup that meets once a week. In the case of linkage with a formal financial institution,the group appoints one of its members to make the monthly payment of its externalloan installment. This feature makes it possible for the SHG’s high-frequency (i.e., weekly)internal accounting and data transmission channels to be conceived independent of physicalcash transfer channels, while the same is not true for the comparatively low-frequency (i.e.,monthly) external dealings of the SHG, where cash transport (and the security issues aroundit) creates a different set of constraints for data management.

In this study, we explore the use of ICTs for capturing and processing information onthousands of microfinance transactions (both internal and external) across hundreds ofremote rural locations in a timely and efficient manner.

4. RESEARCH QUESTIONS AND METHODOLOGY

4.1 Research Questions

We now return to the structuration framework and set up a series of questions about thenature of technology use within the SHGs studied. The objective of these questions is (1) to



act as a set of propositions that will be examined in light of the data collected from the fieldstudy, and (2) to guide the analysis toward possible generalizations. Following Orlikowskyand Robey’s four components or influences in structuration, our central research questionsare as follows:

1. What role did the SHGs have in designing the technology? We assume that the SHGsare the ultimate users of the system because although intermediaries such as thesponsoring organizations are involved, the SHG’s requirements are the target of the“ICT for development” deployment exercise. Here, it is useful to ask how and inwhat manner they were involved with the technology design and selection. How wasthe problem for which ICT support is required recognized and selected?

2. How did the ICT introduction impact the existing processes of lending/accounting?The ICT introduction was made in a space of existing work practices, includingnorms, processes, rules, and institutional constraints. This question focuses on theaspects of change, both accepted and resisted, and the manner in which this changewas wrought.

3. What were the existing procedures of accounting/lending and structural conditionsthat impacted the way SHG members could use technology? Existing procedures andcontexts could either enable or not enable the introduction of ICTs. It is useful toidentify those procedures that help and those enable that do not.

4. How much freedom did the SHGs have in shaping the technology, after implemen-tation, to their needs? How much freedom did they have in redefining their ownprocesses? All SHGs were bound by the ICT implementation and the changes itbrought. A salient issue to examine is how much freedom they had in changing theirown processes and work conditions to effectively use the ICT intervention.

Figure 1 represents the relation between these four key influences.

Figure 1 Applying structuration theory to “ICT for microfinance” interventions.


266 DE’ AND RATAN

4.2 Methodology

In this study, we examine the experiences of two organizations that have attempted to in-troduce ICTs in their front-end microfinance delivery operations. Our criteria for selectionincluded studying microfinance organizations working in different geographic areas, serv-ing a sufficiently large clientele (>100,000 clients), focused on rural operations, having anongoing experiment in the use of technology for development, and with differing choicesof ICT devices to be deployed. We honed in on the NGO PRADAN and the nonbank-ing finance company (NBFC) Sanghamithra, both MFIs that met our various criteria forselection.

We rely on the case study research methodology for this research. The case studymethodology (Lee, 1989; Yin, 2003) provides a rigorous means by which to study thecomplex set of questions that have been raised. A case study “examines a phenomenonin its natural setting, employing multiple methods of data collection to gather informationfrom one or a few entities (people, groups, or organizations)” (Benbasat, Goldstein, &Mead, 1987; p. 370). What is more, a case study is useful when the “boundaries of thephenomenon are not clearly evident at the outset of the research and no experimental controlor manipulation is used” (Benbasat, Goldstein, & Mead, 1987; p. 370). This study wouldalso classify as an interpretive case study, where the context is carefully specified and thepossibility of multiple interpretations and conclusions is retained, along with a fair degreeof “suspicion” (Klein & Myers, 1999; p. 72) toward the data.

We initially collected secondary data on the model developed for the technology interven-tion by each organization. This was complemented with interview data from higher-levelstaff at each organization’s headquarters who were responsible for orchestrating the ICTdeployment across locations. We then collected primary qualitative data from field visitsto five rural deployments across the two MFIs studied, four in Eastern India (West Bengaland Orissa) and one in Southern India (Karnataka), over the course of 4 weeks in August2006. An additional visit was made to one organization’s field site in Central India (MadhyaPradesh) for a week in July 2007. Details on the day-to-day operations of the staff usingthe ICT devices in the field were specifically recorded. At the same time, we interviewedmicrofinance clients of each organization to understand their views on the changes broughtin by the technology intervention. On the whole, data were collected through more than45 hours of open-ended interviews, both in individual and group sessions. We interviewed24 key stakeholders (21 individuals and 3 groups)—from senior MFI management, loan of-ficers, professional field staff, ICT service providers, and bank managers to SHG membersthemselves.

Our analysis involved performing a comparative assessment of the two case studiesat a systems level, overlaying the field-level operational details of the two ICT-enabledmicrofinance delivery systems on their designed “models” to highlight where the two werealigned and where they diverged for each organization. Given that the first MFI we examinedwas a network of SHGs promoted and supported by the NGO PRADAN, the emphasis wasclearly on improving the internal accounting systems of each SHG. In contrast, the secondcase revolved around the network of SHGs borrowing from the NBFC Sanghamithra, and sothe constraints to be targeted were mostly around the external financial transactions of eachSHG. Our evaluation of success in these two types of related microfinance tasks and theuse of ICTs to enable them is, therefore, directly in reference to the operational constraintstargeted by the implementing organization in each case. The degree of divergence betweenthe conceptualized model and the expected benefits ex ante, from the introduction of the



technology channel and the observed ex post real outcomes in the field, is therefore oneof our central indicators of accomplishment. This is corroborated by whether the ICTintervention was scaled beyond the pilot phase and mainstreamed in the MFI’s operations,or abandoned post-pilot implementation.

5. CASE 1: PRADAN’S “COMPUTER MUNSHI” PROJECT

PRADAN is a nonprofit, nongovernmental organization that supports more than 7,500microfinance SHGs across 2,537 villages in 27 districts of 7 states, primarily in India’sCentral and Eastern regions, reaching out to 101,568 low-income households in some ofthe country’s least economically developed rural areas (as of March 31, 2006). PRADAN’sprofessional staff coordinates groups of rural women to form SHGs for saving and borrow-ing, as a first step in organizing rural communities for development activities. A group of 10to 12 SHGs in an area form a “cluster” and 10 to 20 clusters form a “federation” of SHGs.PRADAN then uses this network of SHGs to plan and conduct interventions in livelihoodsupport and promotion. Importantly, PRADAN itself does not operate as a financial in-termediary, with each promoted SHG being responsible for conducting and recording itsown transactions and managing its linkages for external borrowing and saving directly withcommercial banks. PRADAN’s role is simply one of an enabler, monitoring the networksand intervening when there is a problem (Ratan, 2007).

Over the course of interactions with the groups it had promoted, PRADAN was repeat-edly faced with a number of issues around financial record management by the SHGs.First, the internal paper financial records/registers of the SHGs were often incomplete andregularly had major errors in calculation and tallying. This was to be expected because thecomputations required at each weekly meeting were intensive as a result of the flexibilityoffered by the SHGs to their members around interest and principal repayment, as wellas voluntary savings deposits. Even if literate, the chosen SHG group accountants (GAs),whether internal or external, dreaded their accounting task. What this resulted in was ex-tended SHG meetings that took an average of 2 to 2.5 (sometimes up to 4) hours eachweek, mostly involving time spent in doing manual calculations and accounting. Whenaggregated, this has an opportunity cost of losing 8 to 10 hours or 1 full day of work eachmonth per SHG (i.e., around Rs. 360 [∼US$9] in lost wages for a group of 12 membersin a month). The lack of an accurate, reliable, easy-to-prepare, fully updated record of theSHGs’ financial performance hurt the members’ confidence in the microfinance system.It also prevented PRADAN from assessing the viability of the microfinance collectives itwas promoting and identifying problem areas or groups from the records early enough tointervene.

This accounting problem was further compounded during annual auditing, when divi-dends on the group’s pooled capital needed to be calculated and distributed to each SHGmember proportional to her individual savings contribution. The mistakes in weekly entriesby the GAs showed up at this time, and reconciling the internal SHG member accountswith external bank group records and member beliefs/statements about their repaymentwas very difficult, requiring professional auditors’ help and an entire month of dedicatedaccounting work. In the absence of the skill to handle these complex bookkeeping tasksby themselves, the SHGs then became dependent on PRADAN’s professional educatedstaff to manage and check their accounts, clearly taking away from the vision that theywould be independent community organizations after being linked with the formal financial


268 DE’ AND RATAN

sector. However, the arrangement of PRADAN staff becoming the SHGs’ default charteredaccountants was also detrimental to PRADAN’s objectives. PRADAN’s professional fieldstaff ended up spending much of their time checking and correcting the accounting booksof the SHGs, and were left with little time to run projects around livelihood promotion inthese areas through the SHGs.

To respond to these constraints, PRADAN came up with the idea of outsourcing theweekly internal accounting tasks of the SHGs they promoted to a third-party professionalaccountant. This would reduce their (PRADAN’s and the SHGs’) accounting burden andimprove the quality of SHG records. The accountant or “Munshi” had to have a monetaryincentive to provide this service regularly, yet each SHG would not be able to afford a costlyMunshi “service fee.” The pricing problem was addressed by designing an optimal scalefor operations, so one Munshi would service a large number of SHGs (100–200), chargingeach a nominal fee to maintain their records (Rs. 3 per member per month or ∼Rs. 45[US$1.13] per SHG of 15 members per month).

So that the Munshi could handle such a large number of weekly transactions efficiently, hewould have access to a commodity PC on which the electronic records of the SHGs would bestored (hence the name “Computer Munshi” [CM]). This necessitated that the CM be locatedin an urban or semiurban area, with some electricity infrastructure and maintenance support.The CM’s responsibilities would, therefore, include weekly data entry of the handwrittentransaction records of each SHG, checking calculations and tallying figures, updating theelectronic records for each SHG, calculation of new dues/balances, and transmission ofa hard copy of this information to each SHG before the next weekly meeting. He wouldalso supply aggregated monthly financial information on his client SHGs to the promotingagency (PRADAN), as well as to external stakeholders, for a fee.

To transmit this information between the SHGs in the villages and the CM’s locationevery week through a reliable yet low-cost channel, the CM hires a set of peons whoare residents of the villages served. Each peon collects copies of the Regular MeetingTransaction Statement − 1 (RMTS-1) forms of between 30 and 50 SHGs each week (theseare dropped off in local drop boxes by the SHGs once their weekly meeting is done), whichhe then delivers to the CM at his PC kiosk. The peon then returns the processed RMTS-2print-outs reflecting updated balances and dues for the next week to the individual SHGsin his area before their next weekly meeting. The CM pays each peon Rs. 4 per SHG perweek for this data transport work (Ratan, 2007). Figure 2 describes the tasks and scheduleof the various players in the CM system.

The software for this SHG accounting system was developed by a small Delhi-basedservice provider in 2002. It runs on Windows 98 and XP. PRADAN’s “Computer Munshi”project was initially funded by CARE, NABARD (National Bank for Agriculture andRural Development), SIDBI (Small Industries Development Bank of India), and otherdonor agencies, who provided the existing set of computers as grants. Each individualimplementation was handled by the local PRADAN staff in each block/district office androlled out at their own pace. Each branch office was therefore responsible for the hiring ofCMs, training, pricing of services, synchronization of manual and electronic records, andall tasks related to the initialization of the new system. Although the pilot was designed andimplemented in a handful of PRADAN centers starting in 2003, the program has since beenexpanded nationally to all locations where PRADAN works. There are 45 CMs servingPRADAN-promoted SHGs nationwide at present.



Figure 2 PRADAN’s “Computer Munshi” workflow.

6. CASE 2: SANGHAMITHRA’S SIMPUTER PROJECT2

Sanghamithra Rural Financial Services (SRFS) was established by the NGO MYRADA in1995 as an MFI that would both promote and directly lend to rural SHGs, mostly in thesouthern Indian state of Karnataka. SRFS has an outreach of more than 112,000 clientsthrough 9,000 SHGs (2005) and receives capital from a variety of organizations, includingSIDBI, NABARD, and other commercial banks, which it then re-lends to individual SHGs.One interesting aspect of SRFS’ operations is that it does not independently maintainbranches in remote rural locations, but instead piggy-backs on the infrastructure set up byits partner NGO MYRADA and existing public sector banks to conduct its transactions

2This section is based on Ganesan and Pichai (2006).


270 DE’ AND RATAN

Figure 3 Preintervention workflow for loan repayment at SRFS.

with remote rural clients. SRFS therefore uses 40 community-managed resource centers(CMRCs), each run by a federation of SHGs (100–120 SHGs) in a given area, to reach outto its clients, while maintaining its central management information system (MIS) at itsheadquarters in Mysore city. The internal operations of SRFS-linked SHGs are analogousto the SHGs supported by PRADAN.

SRFS faced very high transaction costs when it came to servicing the microcredit it lentout to rural SHGs. To keep costs low, SRFS only had 39 field staff in early 2005, whichmade for a ratio of 2,873 borrowers or roughly 230 SHGs per staff member (Ganesan &Pichai, 2006). With such an extreme servicing ratio, field operations could only be run byhaving the rural SHG members themselves come to some central location to transact withSRFS instead of requiring SRFS’ field officers to visit all clients at their doorstep everymonth. Hence, each rural SHG needed to have one to two of its members visit the localCMRC office once a month to conduct transactions with SRFS (Figure 3). Transport costs(making a roundtrip from their village to the CMRC and back) and the opportunity cost ofmissing 1.5 days of work to make one such loan repayment transaction cost each SHG upto Rs. 80 (∼US$2) in lost wages per month.

In addition, fearing robbery if their field staff and clients carried or held large sums ofmoney in isolated locations, SRFS collected/disbursed all payments from/to its membersthrough checks. This made for slow and costly processes for loan approval, disbursal,and repayment, the brunt of which was borne by the SHG members. New loan disbursal,for instance, involved the SHGs again facing high travel costs and the likely loss of2 day’s wages by the SHG representative, in first collecting the check from the CMRC



and depositing it in the SHG’s bank account in the town, and then returning to the town towithdraw money from the bank on/after the day the check is encashed. Moreover, althoughinterest on the SRFS loan was calculated from the day of approval/disbursal at SRFSheadquarters, the SHG members only received the money several days or weeks later, afterthe check was delivered to the CMRC, collected by the SHG’s representative, depositedand encashed in the SHG’s bank account, and then withdrawn for internal distribution tothe final borrowers (Ganesan & Pichai, 2006).

Given the high transaction costs associated with each monthly loan repayment transactionfor an individual SHG, some preferred to give the loan officer/CMRC staff a number ofpostdated checks at one go, to be deposited over a period of time, so that they would nothave to travel to the bank and the CMRC each month when a repayment was due. However,these checks were for standard equal monthly installment (EMI) amounts; thus, if on theactual day the repayment is due, the SHG wanted to make a payment higher than therequired installment (the excess being paid toward principal so that the loan balance andconsequently future interest was lowered), they could not do so.

A central difficulty expressed by the SHG members was the inability to check theirupdated SRFS loan records, including their transaction history, balance amounts due, andfuture payment schedule. This reduced the transparency of the system, and opened thedoor for fraud/low accountability of the loan officers and CMRC staff who processedthese repayments. In fact, the only records the SHGs possessed of their loan repaymenttransactions were photocopies of their bank check deposit receipts (because the originalreceipt had to be handed in at the CMRC for recording the repayment transaction).

SRFS was additionally burdened by redundant data entry of the same repayment trans-action at several points and the resulting errors in their final records. For instance, theloan repayment transaction was first noted in SRFS’ bank account at the time of checkdeposit by the SHG representative. It was then recorded in a register at the CMRC by thestaff person collecting the repayment check deposit receipt (Figure 3). Finally, once thepaper receipts were collected from the various CMRCs by the SRFS loan officers, the datafrom the receipts was again manually typed into the central MIS at SRFS’ head office inMysore. These multiple rounds of data entry created their own set of errors, reducing theaccuracy of the final records. Furthermore, there were considerable delays between whentransactions occurred in the field and when they were reflected in SRFS’ database, thusmaking it difficult for SRFS to know its exact financial position and cash balance at anypoint in time, which made for poor planning and management of cash flows.

To try and address these problem areas, SRFS decided to pilot an ICT-based front-end system to have data transferred electronically from the point of transaction with theSHG representative directly to the central database maintained at SRFS headquarters. Thecomputing device could travel with the SRFS field officer or sit at the CMRC. When the SHGrepresentative made their loan repayment, the transaction would be recorded and uploaded,in the SHG representative’s presence, by the staff person using the device. Given that datawould be transferred and records updated in real-time, the possibility of error in data entryor fraud would be minimized or caught instantly. Therefore, SRFS envisioned being able tocollect repayments from SHGs in cash at the CMRC in the new system, thereby eliminatingthe time costs involved in the regular cash-check conversion. At the end of the day, the staffperson collecting the payments would be liable for the cash equivalent of all transactionsrecorded in the central database that day. So, although the cash did not travel back andforth, the real-time exchange of accurate financial data on transactions would permit theCMRC to be used almost like a point-of-service banking outlet for SRFS (Figure 4).


272 DE’ AND RATAN

Figure 4 Envisioned Simputer workflow for loan repayment at SRFS.

On the client side, each SHG would own a smart card (complete with a secret PIN)that would act as its electronic passbook. Once the transaction was complete, the SHGrepresentative could have the group’s smart card updated by the device. The SHG couldfurther check the details of not only their current transaction, but also their group’s trans-action history, balance amounts, credit limit, and future repayment schedule by insertingtheir smart card into a smart card reader that would be provided by SRFS to each SHG.Similarly, at the time of disbursing a new loan, instead of the check for the loan amountbeing delivered by the SRFS loan officer to the CMRC, from where the SHG representativewould have to pick it up and deposit it in their bank account, the new system would allowcredit approval for the new loan to be sent over the Internet from SRFS head office to theCMRC. The CMRC would then be able to disburse the loan amount in cash to the SHGrepresentative. This would save both the SRFS field officer’s trip to the CMRC to drop offthe loan checks and the SHG representative’s trip(s) to deposit the loan check and laterwithdraw the amount in cash from the bank.

The device of choice for the new system was the Simputer, a locally developed portablecomputer, run on an open source software platform. This device was seen as an appropriatechoice given its mobility, capability of operating on AA batteries, internal modem for dataupload, internal smart card reader, versatility in offering almost the entire range of featuresas a regular PC, compatibility with the smart card application that was also OS based, a



nonvolatile permanent memory, multilingual support, and the expectation that the solutioncould be customized to SRFS’ requirements by local vendors, given that it was open source.In addition to recording SHG loan transactions, the Simputers were meant to offer otherneed-based information services to those visiting the CMRCs as well. The Simputers weredeployed at five CMRC locations for the pilot phase in 2005. However, the system has notbeen expanded to other SRFS field operations beyond the pilot implementation.

7. ANALYSIS

Despite PRADAN’s intervention centering on the improvement of internal SHG operationsand SRFS’ centering on the improvement of external SHG transactions, there are a numberof parallels between the two cases. The intervention in both cases was introduced at asufficiently aggregated level (i.e., servicing 100–200 SHGs from a single location using asingle ICT device). In both cases, one of the major pain points in the existing set-up was thehigh time costs faced by SHGs in making/managing their regular microfinance transactions.In both cases, the maintenance of an accurate, up-to-date, aggregated financial databasethat drew on data from multiple remote locations was a priority. Finally, in both cases, theattempt was to build trust in an electronic system of financial record management and haverural users feel comfortable accessing and using information in this form.

The clear difference between the two cases was the set of problems associated withcash transport in remote locations that burdened SRFS, but which PRADAN’s CM systemdid not have to worry about. We discuss a number of lessons that can be drawn from acomparison of the experiences of these two organizations, within the framework of thestructuration model of ICT adoption by organizations described previously.

7.1 What Role Did the SHGs Have in Designing the Technology?

In both cases, the technology introduction came as a directive from the sponsoring agency,not from the SHGs. The SHGs were experiencing high transaction costs and inconvenienceowing to the manual nature of the existing processes, and were receptive to the idea;however, they had little role in the initial design of the technology solution provided. Thechoice of the particular hardware, the technology delivery platform, the selection of theoperators, and the training provided to the operators was entirely conducted by PRADANand SRFS. Both agencies were addressing their own needs of scarce and overstretchedresources (i.e., having few field personnel dealing with a large number of SHGs), and as aresult, initiated these technology interventions. A power relation was clearly present in theroles of the sponsoring agencies and the SHGs, and this underscored the technology designand implementation choices.

However, in the PRADAN case, the constraints targeted by the intervention were basedon the difficulties faced by the SHGs in conducting their weekly meetings. A central goalof the ICT deployment was to reduce the accounting burden of individual SHGs throughoutsourcing to a third-party provider like the CM. In contrast, for SRFS, the technologysolution did not incorporate any elements of the constraints faced by the end users—theSHGs. The SHG members were clear about the inconvenience posed to them from havingto visit the CMRC each month to conduct their repayment transaction. However, addressingthis constraint by having the SRFS or CMRC officer visit the SHGs at their location withthe device was not considered a priority in the implementation. As a result, the SHGs were


274 DE’ AND RATAN

mostly unaware that there had been any change at all in their system of transacting with theMFI, even after the Simputer deployment.

7.2 How Did the ICT Introduction Impact the Existing Processesof Lending/Accounting?

In PRADAN’s case, the time spent by the SHGs on accounting tasks each week was cut byhalf, and the quality and completeness of the data improved substantially as a result of theCM system. The SHG members came to rely on and trust the weekly computer-generatedprint-outs they received from the CM containing the relevant figures on balances and duesneeded for that week’s transactions. Compared to the Rs. 360 (US$9) cost of collective timespent at SHG meetings each month in the earlier system, each group’s regular meetingsunder the CM system took half as much time, i.e., an hour each week or 4–5 hours eachmonth [an associated cost of around Rs. 180 (US$4.50) in lost wages]. The financial recordsof most PRADAN-promoted SHGs in a block were now standardized and stored on a singledatabase that was updated weekly. A number of important ratios on the financial health ofindividual SHGs and their members were generated each month (trial and member balancesheets) and given to the groups and to PRADAN for a fee.

PRADAN ended up spending less of its own human resources in servicing the need ofthe SHGs for adequate accounting support. From the organization’s perspective, an acuteconstraint was targeted, and consequently, the greatest gain was for their own staff, whowere now able to spend time doing their “real” work (i.e., promoting and implementingsocial and livelihood enhancement programs with the SHGs).

In contrast, despite SRFS’ ICT intervention, their client SHGs spent the same or agreater amount of time on their microfinance transaction tasks. Travel and opportunitycosts remained at Rs. 80 (US$2) per group per month. There were no changes in transactioncosts for SRFS as well, and they ended up maintaining more than one information system(Simputer and PC-based databases). As a result, redundant work did not decrease and mayeven have increased under the new system, with no integration of the electronic recordsentered by the CMRC staff person on the Simputer with the central database at SRFSheadquarters. The collection of paper receipts from the CMRCs by the SRFS loan officerand the manual entry of these into the central database in Mysore continued, even as theCMRC performed the exact same task on the Simputer at its location. SRFS used theuploaded data from the CMRC Simputer only to verify whether their own manually entereddata was consistent. The Simputer ended up being used in addition to the manual register,as well as the PC system at SRFS headquarters, replacing no part of the earlier set-up.The continuing multiple points of data entry did not allow for any significant reduction inrecording error either (except any gains from the comparison of the two sets of electronicrecords).

Furthermore, transactions were not conducted in real-time as envisioned in the SRFScase and so did not in any way allow for greater transparency or accountability of theagents recording the transaction. The CMRC staff person responsible for this task collectedthe bank receipts as in the old system and entered these data into the Simputer at herconvenience. Because no printed receipts were given to the SHG representative at the pointof transaction, there was no possibility of catching errors at source and reducing recordingerrors through this system.



7.3 What Were the Existing Procedures of Accounting/Lendingand Structural Conditions That Impacted the Way SHG MembersCould Use Technology?

In the PRADAN case, the CM was purposely located in a semiurban area, a choice necessi-tated by the lack of electricity and skilled operators in the more rural areas where the SHGswere located. There was a clear structural reason for locating the new ICT system in suchan area, given the implementers’ experience with the unreliable electricity infrastructure invillages. The SHGs and PRADAN had to then “work around” the distance and communi-cation problem by using a peon/postman on a bicycle for regular data exchange betweenthe CM’s location and his client SHGs’ scattered village locations.

The basic features and capabilities that the SHG members wanted from the technologywere those of accounting calculations and reporting. These elementary accounting taskscould be adequately addressed by simple accounting software on a commodity PC, whichis the choice of technology that PRADAN made. The existing procedures that PRADANfollowed did not require a complicated technology platform or infrastructure.

This is not to say that more complex infrastructure could not be envisaged for sucha situation. Telecommunication equipment such as satellite terminals and coordinationsoftware, or PDAs connecting to a central database through existing land lines, could havebeen adopted (for a price running to millions of rupees) (Kumar, 2004). However, theemphasis in the design was to deploy those technology features and processes that suitedthe client’s location and conditions in the most cost-effective manner (Garg et al., 2005).A PC running a fairly small set of basic applications, with data received and transmittedmanually through a bicycle-enabled peon were most suited for the local conditions. Theycould be scaled upward on demand or down.

In the SRFS case, although the expressed requirement was for the Simputer to be usedin the field, to be “mobile,” it was ultimately placed in a fixed location so it could in factbe functional, given its short battery life. The data transport task ended up having to beperformed manually by the SHG members and the SRFS officers, as in the previous system.Existing local conditions, such as the vast geographic spread, lack of connectivity from thefield, and need for longer battery life owing to long time periods on the road, prevented theuse of the Simputer in the manner in which it was designed. It is interesting to note herethat the Simputer was originally designed for banking purposes, where bank representativeswould carry it to different branches and collect data that would be aggregated at a centraldatabase.3 However, SRFS’ attempt to use it for this exact function was heavily restrictedby the device’s limitations (short battery life) and the environment’s realities (poor ruralconnectivity).

SRFS, as such, had complex and extensive requirements: looking across the range oftasks to be “automated”—mobile computing, networking, printing, data processing, real-time data transfer, cash transfer, integration with peripheral devices, etc. It was not possiblefor the Simputer to provide these capabilities, and that, too, at once.

An important concern for the SHGs in the SRFS case was whether the CMRC couldaccept cash. Given that this was not possible (for monitoring and regulatory reasons), theyhad to retain check-based payments, which removed the only possible benefit for clients.With no change in the possibility that CMRCs would be able to transact in cash, the costsof traveling to the bank to make the cash-check conversion and check deposit remained

3Simputer cofounder Deshpande at Indian Institute of Management Bangalore (IIMB) Seminar, 2006.


276 DE’ AND RATAN

as high as before. With no cash coming into the CMRCs, they could not be expected tohand out cash for new loans even if approval for the same was sent to them electronicallyfrom the SRFS head office. The high transaction costs of SRFS’ field agents (collectingreceipts and dropping off checks from the CMRCs) and SHG clients (traveling to the bankand CMRC to transact each month), therefore, remained unchanged.

7.4 How Much Freedom Did the SHGs Have in Shaping the Technology,After Implementation, to Their Needs? How Much Freedom Did They Havein Redefining Their Own Processes?

In the PRADAN case, SHGs were able to leverage the fact that they paid for the servicesof the CM, and were able to negotiate the fee and payments. The sponsoring agency helpedto set up the CM’s operations and then moved out of the picture, leaving the SHGs withconsiderable scope in arranging for their own data processing and reporting requirements.SHGs were able to obtain services at the quality level they desired and in the time framethat was convenient for them. If the CM’s accounting service failed to meet their needs orwas unsatisfactory, all they would have to do was withhold payment and discontinue theuse of the CM service. This was the benefit of building the ICT intervention as a privatelyprovided and paid service.

The computers initially used in the CM pilot were received as grants to PRADAN,but they were then transferred to the names of the SHG federations using the computer’sservices. In one location, the CM paid the SHG federation for the PC asset he used to runhis business, through monthly installments. Only after he paid for the full amount of the PCwas it transferred to him as private property. This ensured that in the interim (1−2 years),the CM was accountable downward to the collective that owned the asset he used (who alsohappened to be the customers he served).

There were two aspects of the innovation introduced by PRADAN that ensured itsscalability. First, the CM in several locations was an independent entity who was free tooffer other value-based services to sustain her business. Second, the CM, a human entity,used the ICT tool to enable her effectiveness, thus ensuring that in the future, based on therequirements of the market, she would continue to innovate and use necessary technology toprovide other services, perhaps even to other interested parties such as banks and financialinstitutions interested in delivering products and services to the SHG members.

In the SRFS case, the Simputer device was at the heart of the innovation introduced bySRFS. It found few complements in the users who could adapt the technology to the needsof the situation. The user interface was not intuitive, for example, and the users made noattempts to adapt it to their own situation, for instance, by practicing on the device so thatthey could enter data fluidly.

The SHG users had expressed very clearly defined needs for which they needed support—being serviced at their location, being able to see particular fields of aggregated information,etc. However, these were not addressed by the implemented ICT system. This prevented theSHGs from appropriating the technology and redefining it according to their own needs.For example, the system did not show the SHG’s “loan balance” and “future repaymentschedule,” which were the two essential pieces of data end users looked for in such aclient-facing information system. The users did request the technology vendor to makethis change. The vendors who designed the software were contacted a number of times,but they were not responsive and showed low commitment to the project (in the face of



more profitable work coming their way). Because the vendors had been contracted by thesponsoring agency, the SHGs could exercise no power over them for continued support.

Neither the persons who used the Simputer at the CMRCs nor those for whom it wasused were aware of the ways in which the device was supposed to help them in their work.Furthermore, the smart cards for the SHGs ended up being kept at the CMRC itself, withthe CMRC staff person entering the passwords both for herself and the SHG representative.In fact, many SHGs had no idea that a device like the Simputer existed at the CMRC wherethey conducted their transactions. This lack of awareness and ownership of the technologyby those who used it made for weak relationships of accountability and, consequently, poorunimaginative use of the technology.

8. CONCLUSIONS

Through an assessment of the field experiences of two organizations that tried to improvetheir field-level microfinance operations using ICTs, we have shown how the process oftranslating theoretical ICT capabilities into real and measurable improvements for the ruralpoor is far from automatic. Applying the structuration model of information technologyto the instances described has shown how human actors, technology, and the institutionalenvironment interact in particular ways to produce deployments of ICTs for developmentthat are more or less effective in achieving their stated goals. Central to the usefulness ofany technology deployment in this space is the need to define the parameters of “efficiency”and their enhancement in direct relation to the end beneficiary, in this case, the microfi-nance customer. Where efficiency targets are devoid of client pain points, the resultingdeployments are neither useful nor sustainable or scalable.

The value of comparative studies such as ours is not to pat one organization on theback and chide another, but indeed to point these and other organizations to the generalprinciples learned from individual experiences. It is our contention that the success ofany organization (including the ones studied) in solving ICT-enabled service provisionproblems in rural settings will be determined by how their ICT implementation interactswith various organizational features and choices that produce outcomes closer to the client’sconstraints than not. In this sense, the decision-making processes around ICT deploymentsassessed here can be expected to be relevant to microfinance operations not only in otherregions of India, but in other low-income settings in Asia and Africa as well where thevast majority of microfinance activity occurs. The infrastructural, cost, and institutionalcontexts in areas with higher average incomes and greater individual lending such as LatinAmerica may differ in significant ways that may not allow for the hybrid technologicalsolutions described here to be as relevant.

Specifically, the two case studies presented in this article highlight the contextual andpolitical issues pertinent to the success of an ICT for development implementation. Contextis defined by the technology, the human actors who deploy and use the technology, and thesocioeconomic setting in which the situation unfolds. Although the two case studies arebroadly similar in their mission and comparable at a macro level, the contexts and mannerin which ICT was introduced in each varied and led to starkly different consequences.While in the PRADAN case, the discussion of efficiency enhancements drew heavily fromthe microfinance client’s requirements, the SRFS case involved a front-end deployment inwhich clients’ stated needs were sidelined in favor of organizational convenience, leavingthe clients almost entirely unaware of and unaffected by the ICT deployment.


278 DE’ AND RATAN

Our results resonate with findings of earlier research in that “appropriate technology”is that which aligns with locally relevant practices (Avgerou, 1998, 2001). Even thougha device such as the Simputer, explicitly developed for Indian conditions, was used inthe SRFS case, it was still not appropriate because it was not suitable or adaptable tothe peculiarities of usage, prevailing customs, stock of skills, and available resources forcontinued use in the given rural setting. In contrast, a commodity technology such as aPC running basic accounting software was successful, primarily because it could be easilymolded to the needs and constraints of the prevailing conditions through a client-centereddecision-making process.

The Simputer’s business model has been criticized for not enabling local users and en-trepreneurs to use and enhance its special features (Chesbrough et al., 2006). The Simputersignified an “appropriate technology” for Indian conditions by reason of its provenance,and this enabled its adoption by SRFS (as well as other development agencies). However,the Simputer developers had to partner with the adoptive users and build its capacities, andhelp it adapt to local and particular conditions, which they did not do. It is important tonote here that the technology choice was made by SRFS and not the SHGs. SRFS was in aposition to exercise influence over its vendor, the Simputer makers, which the SHGs couldnot, but SRFS did not do so. The choices of human actors around communication, action,conflict resolution, and reward are the lifeblood of strong organizational functioning andchange, and technology deployments are not immune to the influence of these forces.

One of the reasons why PRADAN and its partner SHGs were able to gain from technologyuse, at a broad level of understanding, is that they were able to obtain positive results froma tightly defined and limited application (i.e., they were able to achieve first-order effects)(Malone & Rockart, 1991). We conclude that it is important to achieve such first-ordereffects, and plan to achieve them, as they form the basis for a sustainable technologyadoption process. Users were able to see the value of the technology quickly, to the extentthat they were even willing to pay for the ICT service on a regular basis. Furthermore, usersare then also able to imagine how they can create adaptations that can scale the technologyfor their larger requirements. Second-order effects are possible and are usually the mostvaluable for any organization, only when positive first-order effects are achieved.

The most important first-order effects for rural MFIs are those of reducing client trans-actions costs, which could be achieved by enabling bundling of operations—such as manySHGs using a single accountant, a single SHG submitting multiple checks in one go, ormany cash disbursements being collected in one turn. Bundling reduces the cost of traveland communication, whether these are done manually or by moving information electron-ically. We believe that future ICT initiatives in this particular application domain shouldfocus on enabling such bundling of transactions to lower costs, but only as per the con-venience of clients. As in the SRFS case, enabling bundling by allowing transaction coststo be transferred entirely to poor clients, despite the ICT intervention, might reduce coststo the organization and enhance efficiency, but the development goal of the microfinanceproject would be left unaddressed.

The possibility of transferring remote data through cheap electronic devices and com-munication links remains a promising route that needs to be explored by future research.The need for accurate and up-to-date information is acutely felt by all stakeholders, andICT devices are the most suited to accomplish this task. What our cases illustrate is thatthe capabilities of the technology in and of itself are of limited value if its use is in aone-off “proof of concept” instance. Organically integrating the new device and the newwork processes into the existing routines of organizational exchange with its clients, in the



context of infrastructural limitations, cost considerations, and customer preferences, is whatmakes or breaks a successful and sustained deployment of information and communicationtechnologies for development. We hope that the effectiveness of client-centered contextualinnovation processes highlighted in this study is kept center stage as discussions aroundnew ICT deployments, especially involving the marriage of mobile payment systems andmicrofinance operations, gain steam.

9. CONTRIBUTIONS AND FUTURE WORK

The main contributions of the research reported in this article are identifying a frameworkfor the application of structuration theory in the ICTs for development domain, highlight-ing the political nature of defining and designing “efficiency-enhancing” technologies indevelopment contexts, and describing the importance of how technology introductions aremanaged for better or worse by human agents.

First, structuration provides a powerful theoretical basis for asking questions about tech-nology introductions for development. This is a well-established theory for understandingcomplex phenomena, and this article extends earlier research by identifying a framework,or a set of research questions, that may be used in an ICT for development context. Also,prior research has not included the use of structuration in an application of ICT use inmicrofinance agencies.

Second, we provide evidence for technology interventions as inherently political pro-cesses, involving marked choices and trade-offs between catering to a customer’s contextand needs and providing organizational convenience. The manner of deploying a giventechnology itself becomes an expression of political choices. General-purpose computingdevices that can be configured for local and context-specific needs are useful under certainsituations—the most important of which is the flexibility that the users have in shapingthe device and the processes of using it to their current and emergent needs. The articlecontributes to the literature by showing two contrasting and detailed cases, one in whichthe technology was shaped to the needs of the clients and another in which such action wasnot undertaken.

Prior research that examines the nature of human agency in introducing technology withinan organization has been sparse. Our research shows that technology introductions may beinitiated by a powerful, external agent. However, it is imperative that end users’ constraintsbe the ones targeted and that they are able to appropriate the technology, with the helpof the external agent. Where the microfinance organization’s staff pursues the clients’ keyproblems in their given context as the core issues to be streamlined and made efficient, thetechnology deployment is able to achieve development impact. Otherwise, the deploymentonly adds to the list of technology for development projects that ride on hype around thepromise of advanced technological capabilities and fold after a brief experimentation phase.

The limitations of this research point to the future work that needs to be done. The currentwork focuses on ICT introduction in the context of microfinance. This context has to bebroadened in the future to consider other domains in which ICTs are introduced and wherethe research questions we have considered are applicable. Two concrete examples of suchcontexts are the use of ICT in microenterprises, particularly those engaged in by the urbanpoor and the use of ICT in rural mass education. Another limitation of this research is thatit considers only two MFIs, a limitation that will be overcome in the future by consideringdifferently enabled and differently located institutions.


280 DE’ AND RATAN

ACKNOWLEDGMENTS

Thanks to Kentaro Toyama for his comments on an earlier draft, to Hariprasad Pichai andSrikrishnan Ganesan for their help with data collection, and to Saleela Patkar for someuseful insights. Our research would not have been possible without the cooperation of theorganizations studied—many thanks to the staff and members of SRFS and PRADAN.

GLOSSARY OF ACRONYMS

CM Computer MunshiCMRC community-managed resource centerEMI equal monthly installmentGA group accountantICT information and communication technologiesMFI microfinance institutionMIS management information systemMYRADA Mysore Resettlement and Development AgencyNABARD National Bank for Agriculture and Rural DevelopmentNBFC nonbanking finance companyNGO nongovernmental organizationPRADAN Professional Assistance for Development ActionRMTS Regular Meeting Transaction StatementSHG self-help groupSIDBI Small Industries Development Bank of IndiaSRFS Sanghamithra Rural Financial Services

REFERENCES

Ananth, B., Barooah, B., Ruchismita, R., & Bhatnagar, A. (2004). A blueprint for the delivery ofcomprehensive financial services to the poor in India. Unpublished report, ICICI Bank SocialInitiatives Group, Mumbai.

Armendariz, B., & Morduch, J. (2005). The economics of microfinance. Cambridge, MA: MIT Press.Avgerou, C. (1998). How can IT enable economic growth in developing countries? Information

Technology for Development, 8(1), 15–28.Avgerou, C. (2001). The significance of context in information systems and organizational change.

Information Systems Journal, 11, 43–63.Bailur, S. (2007). The challenges of community participation in rural information systems projects.

IFIP 9.4 Working Group Conference on Social Implications of Computers in Developing Coun-tries: Taking Stock of e-Development, Sao Paulo, Brazil, May 28–30.

Benbasat, I., Goldstein, D., & Mead, M. (1987). The case research strategy in studies of informationsystems. MIS Quarterly, 11(3), 369–385.

Bhatt, N., & Tang, S.-Y. (1998). The problem of transaction costs in group-based microlending: Aninstitutional perspective. World Development, 26(4), 623–637.

Chesbrough, H., Ahern, S., Finn, M., & Guerraz, S. (2006). Business models for technology in thedeveloping world: The role of non-governmental organizations. California Management Review,48(3), 48–61.

De’, R. (2005). E-government systems in developing countries: Stakeholders and conflict. Proceedingsof Electronic Government, 4th International Conference, EGOV 2005, Copenhagen, Denmark,August 22–26.



DeSanctis, G., & Poole, M. S. (2004). Capturing the complexity in advanced technology use: Adaptivestructuration theory. Organization Science, 5(2), 121–147.

Emran, M. S., Morshed, A. K. M. M. & Stiglitz, J. E. (2007). Microfinance and missingmarkets. Working paper series. Retrieved August 18, 2009, from http://papers.ssrn.com/sol3/papers.cfm?abstract id=1001309

Frankiewicz, C. (2003). Information technology as a strategic tool for microfinance in Africa: Aseminar report. Paper presented at AfriCap: The First African Investment Fund for CommercialMicrofinance, Nairobi, Kenya, April 26–27.

Ganesan, S., & Pichai, H. (2006). Study of delivery of financial services in rural areas through ICT.CCS Report, Indian Institute of Management Bangalore (IIMB).

Garg, N., Sobti, S., Lai, J., Zheng, F., Li, K., Krishnamurthy, A., & Wang, R. (2005). Bridging thedigital divide: Storage media + postal network = generic high-bandwidth communication. ACMTransactions on Storage (TOS), 1(2), 246–275.

Giddens, A. (1984). The constitution of society. Cambridge, UK: Polity Press.Gurau, C. (2005). ICT strategies for development: Implementing multichannel banking in Romania.

Information Technology for Development, 11(4), 343–362.Haqqani, A. B. (2003). The role of information and communication technologies in global develop-

ment: Analyses and policy recommendations. Report of United Nations ICT Task Force.Ivatury, G. (2006). Using technology to build inclusive financial systems. Focus Note No. 32, Con-

sultative Group to Assist the Poor (CGAP), Washington DC.Jones, M. R., & Karsten, H. (2008). Gidden’s structuration theory and information systems research.

MIS Quarterly, 32(1), 127–157.Kamel, S. (2005). The use of information technology to transform the banking sector in developing

nations: Editorial introduction. Information Technology for Development, 11(4), 305–312.Kannabiran, G., & Narayan, P. C. (2005). Deploying Internet banking and e-commerce—Case study

of a private-sector bank in India. Information Technology for Development, 11(4), 363–379.Keniston, K. (2003). IT for the masses: Hype or hope? Working paper. Retrieved May, 2006, from

http://www.mit.edu/people/kken/PAPERS/EPW paper.htmlKhandker, S. R. (2005). Microfinance and poverty: Evidence using panel data from Bangladesh.

World Bank Economic Review, 19(2), 263–286.Kiri, K., & Menon, D. (n.d.). For profit rural kiosks in India: Achievements and challenges. Informa-

tion for Development (i4d) Online. Retrieved August 18, 2009, from www.i4donline.net/articles/current-article.asp?articleid=700&typ=Features

Klein, H. K., & Myers, M. (1999). A set of principles for conducting and evaluating interpretive fieldstudies in information systems. MIS Quarterly, 23(1), 67–94.

Krishna, S., & Madon, S. (Eds.). (2003). Preface to the digital challenge: Information technology inthe development context. Aldershot, UK: Ashgate.

Krishna, S., & Walsham, G. (2005). Implementing public information system in developing countries:Learning from a success story. Information Technology for Development, 11(2), 123–140.

Kumar, R. (2004). E-choupals: A study on the financial sustainability of village Internet cen-ters in rural Madhya Pradesh. Information Technologies and International Development, 2(1),45–73.

Kuriyan, R., Ray, I., & Toyama, K. (2006). Integrating social development and financial sustainabil-ity: The challenges of rural kiosks in Kerala. Proceedings of the 1st ACM/IEEE InternationalConference on ICT and Development, Berkeley, CA, May 25–26.

Lee, A. S. (1989). A scientific methodology for MIS case studies. MIS Quarterly, 13(1), 33–50.Malone, T. W., & Rockart, J. F. (1991). Computers, networks and the corporation. Scientific American

(Special Issue), 6(1), 140–147.Mathison, S. (2005). Increasing the outreach and sustainability of microfinance through ICT in-

novation. In S. Mathison (Ed.), Electronic banking with the poor: Increasing the outreach andsustainability of microfinance through ICT innovation (pp. 1–7) Brisbane, Australia: The Foun-dation for Development Cooperation.

Mauguis, V., Choucri, N., Madnick, S. E., Siegel, M. D., Gillett, S. E., Haghseta, F., Zhu, H., & Best,M. L. (2005). Global e-readiness—For what? Readiness for e-banking. Information Technologyfor Development, 11(4), 313–342.

Miscione, G. (2007). Telemedicine in the upper Amazon: Interplay with local health care practices.MIS Quarterly, 31(2), 403–425.


282 DE’ AND RATAN

Orlikowski, W. J. (2000). Using technology and constituting structures: A practice lens for studyingtechnology in organizations. Organization Science, 11(4), 404–428.

Orlikowski, W., & Robey, D. (1991). Information technology and the structuring of organizations.Information Systems Research, 2(2), 143–169.

Orlikowski, W., & Scott, S. V. (2008). The entangling of technology and work in organizations.Working Paper Series No. 168, Department of Management, Information Systems and InnovationGroup, London School of Economics and Political Science, London.

Parikh, T., Javid, P., Sasikumar, K., & Ghosh, K. (2006). Mobile phones and paper documents:Evaluating a new approach for capturing microfinance data in rural India. Proceedings of CHI,Montreal, Canada, April 22–27.

Parikh, T. S. (2006). Rural microfinance service delivery: Gaps, inefficiencies and emerging solutions.Proceedings of the International Conference on ICT and Development, Berkeley, CA, May 25–26.

Parthasarathy, B. (2004). Information and communication technologies for development: A compara-tive analysis of impacts and costs from India. Project report, International Institute of InformationTechnology – Bangalore (iiit-b).

Prakash, A., & De’, R. (2007). Importance of development context in ICT4D projects: A study ofcomputerization of land records in India. Information Technology and People, 20(3), 262–281.

Press, L. (2005). Toward a global rural network: Strategy and action plan. Information Technologyfor Development, 11(4), 399–405.

Punnathara, C. J. (2007). “Accredited loan providers—A placebo for rural India.” The Hindu,August 4. Retrieved August 17, 2009, from www.thehindubusinessline.com/2007/08/04/stories/2007080450120800.htm

Puri, S. K. (2007). Integrating scientific with indigenous knowledge: Constructing knowledge al-liances for land management in India. MIS Quarterly, 31(2), 355–379.

Rangaswamy, N. (2006). Social entrepreneurship as critical agency: A study of rural Internet kiosks.Proceedings of the 1st ACM/IEEE International Conference on ICT and Development, Berkeley,CA, May 25–26.

Ratan, A. L. (2007). An assessment of PRADAN’s “Computer Munshi” intervention to improvemicrofinance accounting operations. Microsoft Research India Case Study, September.

Regy, P. V., & Mahajan, V. (2006). IT@BASIX successes and failures: A retrospective. Informationfor Development (i4d) Online, May, 8–10. Retrieved August 18, 2009, from www.i4donline.net/May06/665.pdf

Stone, D. (2002). Policy paradox: The art of political decision making. New York: W. W. Norton.Walsham, G., Robey, D., & Sahay, S. (2007). Foreword: Special issue on information systems in

developing countries. MIS Quarterly, 31(2), 317–326.Walsham, G., & Sahay, S. (2005). Research on information systems in developing countries: Current

landscape and future prospects. Information Technology for Development, 12(1), 7–24.Yin, R. K. (2003). Case study research: Design and methods (Vol. 5). Thousand Oaks, CA: Sage.

Rahul De’ is the Hewlett-Packard Chair Professor in ICT for Sustainable Economic Development atthe Indian Institute of Management Bangalore. He has a PhD (1993) from the Katz Graduate Schoolof Business of the University of Pittsburgh, an MBA (1985) from University of Delhi, and a BTech(1983) from IIT Delhi. His current research interests focus on developing countries, and includeissues such as ICT design, e-governance impacts and evaluation, and the economic impact of opensource software.

Aishwarya Lakshmi Ratan is an Associate Researcher with the Technology for Emerging MarketsGroup at Microsoft Research India. She has a Master’s degree in Public Administration and Interna-tional Development from the Kennedy School of Government, Harvard University, and a Bachelor’sdegree in Economics from Wellesley College in Massachusetts. Her current research involves studyingappropriate technology interventions to enhance microfinance delivery, mobile phone–enabled bank-ing, drivers of household economic mobility, and the effects of IT skills acquisition on low-incomeworkers’ welfare.


Whose gain is it anyway? Structurational perspectives on deploying ICTs for development in India's...

Documents

Transcript of Whose gain is it anyway? Structurational perspectives on deploying ICTs for development in India's...