ArticlePDF Available

A Matter of (Relational) Style: Loan Officer Consistency and Exchange Continuity in Microfinance

September 2015
Management Science 62(4)

DOI:10.1287/mnsc.2015.2167

Authors:

Jason Greenberg

Cornell University

Rodrigo Canales

Boston University

Social scientists have long considered what mechanisms underlie repeated exchange. Three mechanisms have garnered the majority of this attention: formal contracts, relational contracts, and relationally embedded social ties. Although each mechanism has its virtues, all three exhibit a common limitation: an inability to fully explain the continuation and stability of intertemporal exchange between individuals and organizations in the face of change. Drawing on extensive quantitative data on approximately 450,000 microfinance loans made by a microfinance institution in Mexico from 2004 to 2008 that include random assignment of loan officers, this research proposes the concept of ”relational styles” to help explain how repeated exchange is possible in the face of personnel change. We define relational styles as systematically reoccurring patterns of interaction employed by social actors within and across exchange relationships—in this paper, between microfinance clients and loan officers. We show that relational styles that are consistent facilitate a clear understanding of expectations and thus exchange. We also demonstrate that consistency in the relational styles followed by successive loan officers mitigates the negative impact of a broken loan officer–client tie. This paper thus proposes and empirically tests a social mechanism based on relational styles that often accompanies relational embeddedness, but which may also serve as a partial substitute for it.

A Change in Loan Officer Results in a Higher Rate of Missed Payments

…

Missed Payment Rates Vary by Loan Officer Relational Enforcement Styles

…

Missed Payment Rates Vary by Transitions Between and Within Loan Officer Relational Enforcement Styles

…

Figures - uploaded by Jason Greenberg

Content may be subject to copyright.

Content uploaded by Jason Greenberg

Content may be subject to copyright.

This article was downloaded by: [128.122.183.209] On: 09 September 2015, At: 12:12

Publisher: Institute for Operations Research and the Management Sciences (INFORMS)

INFORMS is located in Maryland, USA

Management Science

Publication details, including instructions for authors and subscription information:

http://pubsonline.informs.org

A Matter of (Relational) Style: Loan Officer Consistency

and Exchange Continuity in Microfinance

Rodrigo Canales, Jason Greenberg

To cite this article:

Rodrigo Canales, Jason Greenberg (2015) A Matter of (Relational) Style: Loan Officer Consistency and Exchange Continuity in

Microfinance. Management Science

Published online in Articles in Advance 03 Sep 2015

. http://dx.doi.org/10.1287/mnsc.2015.2167

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial use

or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher

approval, unless otherwise noted. For more information, contact permissions@informs.org.

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness

for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or

inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or

support of claims made of that product, publication, or service.

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, management

science, and analytics.

For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org

MANAGEMENT SCIENCE

Articles in Advance, pp. 1–23

ISSN 0025-1909 (print) ISSN 1526-5501 (online) http://dx.doi.org/10.1287/mnsc.2015.2167

A Matter of (Relational) Style: Loan Ofﬁcer

Consistency and Exchange Continuity in Microﬁnance

Rodrigo Canales

Yale School of Management, New Haven, Connecticut 06520, rodrigo.canales@yale.edu

Jason Greenberg

Leonard N. Stern School of Business, New York University, New York, New York 10012, jgreenbe@stern.nyu.edu

Social scientists have long considered what mechanisms underlie repeated exchange. Three mechanisms have

garnered the majority of this attention: formal contracts, relational contracts, and relationally embedded

social ties. Although each mechanism has its virtues, all three exhibit a common limitation: an inability to fully

explain the continuation and stability of intertemporal exchange between individuals and organizations in the

face of change. Drawing on extensive quantitative data on approximately 450,000 microﬁnance loans made by

a microﬁnance institution in Mexico from 2004 to 2008 that include random assignment of loan ofﬁcers, this

research proposes the concept of ”relational styles” to help explain how repeated exchange is possible in the face

of personnel change. We deﬁne relational styles as systematically reoccurring patterns of interaction employed

by social actors within and across exchange relationships—in this paper, between microﬁnance clients and loan

ofﬁcers. We show that relational styles that are consistent facilitate a clear understanding of expectations and

thus exchange. We also demonstrate that consistency in the relational styles followed by successive loan ofﬁcers

mitigates the negative impact of a broken loan ofﬁcer–client tie. This paper thus proposes and empirically tests

a social mechanism based on relational styles that often accompanies relational embeddedness, but which may

also serve as a partial substitute for it.

Keywords: economic sociology; organizational sociology; relational sociology; sociology of development; social

networks

History : Received October 8, 2013; accepted January 13, 2015, by Jesper Sørensen, organizations. Published

online in Articles in Advance.

1. Introduction

Formal contracts are fundamental features of eco-

nomic and social life. Because they specify rights

and responsibilities ex ante, they facilitate exchange

between actors by reducing the risk of defection.

In complex settings, however, it is impossible to

anticipate the myriad contingencies that may arise,

rendering formal contracts incomplete, difﬁcult to

craft, and costly to enforce (Macneil 1978,Williamson

1985). Relational contracts can minimize these lim-

itations through more ﬂexible structures based on

the trust that develops between parties with an eco-

nomic interest in repeated exchange (the “shadow

of the future”). Ultimately, relational contracts derive

value from relationships between social actors. Tradi-

tionally, these relationships are theorized and docu-

mented between organizations or organizations and

speciﬁc employees or clients (e.g., Rousseau 1990,

McMillan and Woodruff 1999,Baker et al. 2002).

However, organizations cannot establish informal

relationships—only people can (Sorenson and Rogan

2014). Relational contracts are established between

people who set, interpret, enact, and enforce them.

With this as a starting point, sociologists refer to

“relational embeddedness” as the quality of dyadic

relationships that develop between actors over time

(e.g., Granovetter 1992,Heimer 1992,Gulati 1998,

Nahapiet and Ghoshal 1998,Moran 2005). Consistent

with relational contracting, substantial research has

demonstrated the value of relational embeddedness

for individuals and the organizations they represent

(e.g., Uzzi 1997,Uzzi and Lancaster 2003).

A corollary of conceptualizing relationships at the

dyadic level is that when relationships change or

dissolve, so should the value they create for indi-

viduals or the organizations they represent in an

exchange (Seabright et al. 1992,Beatty et al. 1996,

Baker et al. 1998,Broschak 2004,Biong and Ulvnes

2011,Broschak and Block 2014; see also Burt 2000,

2001). This is particularly important when only one

representative of an organization controls the client

relationship. Given the ubiquity of change, employee

turnover, and broken ties, it is unclear how organi-

zations can consistently retain the value of the inter-

personal relationships established by its members

to sustain repeated exchange (Sorenson and Rogan

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

2014). Yet many do to good effect. Our paper provides

one explanation for this puzzle.

We do so using microﬁnance as a strategic research

site. Microﬁnance provides small loans using sim-

ple contracts between microﬁnance institutions (MFIs)

and their borrowers (e.g., Giné et al. 2010). The

frequency of loan disbursements allows for the

observation of a large number of simple contractual

interactions between an organization, its representa-

tives, and its clients. Microﬁnance clients tend to be

destitute and geographically dispersed, so loan ofﬁ-

cers are often their only point of contact with the

MFI. Because client needs vary considerably across

individuals and settings, ex ante contracts are nec-

essarily incomplete, and borrower–loan ofﬁcer rela-

tional considerations carry particular importance. In

essence, microﬁnance constitutes an extreme version

of small business ﬁnance where decentralized banks

have been shown to be more effective because their

branch managers have the discretion and incentives

to establish relational contracts with their clients

(Petersen and Rajan 1994,Berger et al. 2001). Given

that bank employees are routinely promoted, ﬁred, or

rotated, the ensuing broken ties should produce dis-

ruptions that impact continuity of exchange. Notwith-

standing constant change, these decentralized banks

retain an organizational advantage through their rela-

tional contracts.

We determine a way they do so, using a novel,

proprietary data set that includes information on

approximately 450,000 microﬁnance loans made by

an urban-market-focused MFI in Mexico between

2004 and 2008. Theory generation, construct devel-

opment and validation, and model speciﬁcation and

interpretation are aided by rich ethnographic data,

including 129 interviews, collected as part of a larger

research project (Canales 2011,2014). We gain empir-

ical traction from the natural variation in agents’

relational styles and the ﬁrm’s policy of randomly

assigning and rotating loan ofﬁcers across branches

when joining the ﬁrm and in response to vacant posi-

tions created by frequent turnover. The data include

ﬁne-grained information concerning the terms of the

loan, the borrower’s characteristics, and unique mea-

sures characterizing the consistency of each loan ofﬁ-

cer’s relational style, where some agents follow a

strictly contractual approach; others adhere to a holis-

tic, broad interpretation of contractual terms and

client conditions; and others mix elements of both

approaches. Our identiﬁcation strategy exploits the

fact that regardless of why a tie is severed between

a client and the loan ofﬁcer representing the MFI,

the subsequent loan ofﬁcer is assigned at random to

serve the client on behalf of the MFI. This provides

random variation with respect to the (in)consistency

in the previous and subsequent loan ofﬁcers’ rela-

tional styles—orthogonal to client and loan ofﬁcer

characteristics or vacancy—that affords the analytical

leverage required to test our predictions concerning

(in)consistency in inter-loan ofﬁcer relational styles.

We provide compelling evidence that although for-

mal contracts bind the borrower to the MFI, these

formal contracts are enhanced, as expected by rela-

tional contracts and relational embeddedness, by the

loan ofﬁcer’s relationship with the borrower. At the

same time, we show that relational value is not

solely derived from the nature or strength of the

dyadic relationship as characterized by the litera-

ture on relational embeddedness. Rather, clients care

considerably about the relational styles employed by

their previous and succeeding loan ofﬁcers. We the-

orize these relational styles as systematically reoccur-

ring modes of interaction and underlying schemata

and scripts enacted by social actors. Below we pro-

pose theory that speciﬁes how both a loan ofﬁcer’s

individual consistency and the consistency between

successive loan ofﬁcers’ relational styles inﬂuence a

borrower’s adherence to contractual provisions con-

cerning timely repayment.

Our ﬁndings show that when loan ofﬁcers leave the

MFI, which occurs quite frequently both in the ﬁrm

we study and the industry at large (Janik 2012), clients

are approximately 24% more likely to miss a payment,

and contingent on a ﬁrst missed payment, 47% of

clients also miss a second. We show that these ﬁgures

can be reduced signiﬁcantly and rapidly depending

on the consistency of the relational style employed by

the loan ofﬁcer subsequently (randomly) assigned to

administer the loan.

These ﬁndings have considerable economic impli-

cations. MFIs must maintain capital reserves equal to

speciﬁc percentages of their outstanding portfolios at

risk. As clients miss more payments, capital reserve

requirements increase nonlinearly (e.g., from 4% of

the loan value if only one payment has been missed

(1–7 days in arrears) to 30% if three scheduled pay-

ments are missed (30–60 days in arrears)). For an MFI

with a $100 million portfolio (which the MFI we study

has), capital reserve requirements can vary by mil-

lions of dollars.1

The focus on relational styles adds depth to our

understanding of relational contracts (see, e.g., Baker

et al. 2002,Gibbons and Henderson 2011) on two

separate levels. First, it speciﬁes that relational con-

tracts are established through individuals and that,

1Because MFIs leverage capital (8×for conservative ones), the

difference in capital reserve requirements implies the inability to

invest several million dollars. With yearly turnover of invested cap-

ital averaging 3×, and high interest rates charged to clients (mean =

800122, SD =60475), this implies foregone income of millions of

dollars.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

as a result, the dyadic relationship is a core con-

duit of relational value attributed to relational con-

tracts vis-à-vis formal contracts. At the same time, our

research shows that relational contracts and relational

embeddedness do not provide organizations with sus-

tainable value solely through personal ties between

employees and clients. Rather, the consistency in the

styles of interaction employed by social actors occu-

pying similar roles (loan ofﬁcers in this case) provides

a stable and clear set of expectations and thus shapes

action. We demonstrate the value of this social mech-

anism net of the strength and quality of personal ties.

Prior work that focuses solely on the nature and qual-

ity of dyadic ties without observing relational styles

across different individuals may thus conﬂate dis-

tinct social processes—that pertaining to the quality of

a dyadic relationship (relational embeddedness) and

that relating to the particular style used by different

actors as well as the expectations that such styles and

their consistent application can foster.

In the following section, we provide additional

background on microﬁnance and its suitability as a

research setting. We then offer a summary of formal

contracts, relational contracts, and relational embed-

dedness to provide more detail about their virtues

and common limitation—the inability to fully explain

the continuity of exchange in the face of interper-

sonal change. In the theory section, we develop our

arguments and hypotheses specifying how, when, and

why consistency in relational styles of loan ofﬁcers

should matter net of a contractual or dyadic relation-

ship with a client. We then discuss the quantitative

and qualitative data collected, as well as our analytic

strategy. This is followed by a presentation of ﬁndings

and a discussion that includes consideration of scope

conditions and avenues for future work.

1.1. Microﬁnance as a Setting to Research

(Relational) Contracts

Each year, microﬁnance provides approximately

$85 billion in loans to more than 150 million low-

income borrowers (Daley-Harris 2009,Reille and

Forster 2008). Broadly speaking, microﬁnance pro-

vides ﬁnancial services, mostly in the form of

microcredit, to unbanked and often destitute pop-

ulations. Loans are typically small, uncollateralized,

provided for short terms (between four and six

months), and amortized through high-frequency pay-

ments (for a good introduction, see Morduch and

Armendáriz de Aghion 2005). Loans are of the sim-

plest form with terms, ﬁxed rates, and a straight

amortization schedule speciﬁed in a simple legal

document.

From its origins in the 1970s, microﬁnance has

demonstrated that poorer households not only con-

stitute reasonable credit risk but also can put loans

to productive use, even using them to mitigate the

effects of poverty (Yunus 2003). One of the most

remarkable aspects of microﬁnance is that, contrary to

conventional contract and ﬁnance theory (e.g., Bester

1985,Stiglitz and Weiss 1986), destitute populations

with little collateral and no experience with formal

ﬁnance exhibit exemplary repayment rates (Morduch

1994,Morduch and Armendáriz de Aghion 2005). Ini-

tial explanations for this puzzle centered on the com-

mon methodology of providing loans to joint liability

groups where members have the incentive to screen,

monitor, and enforce repayment of joint liability loans

(e.g., Stiglitz 1990, Besley and Coate 1995). However,

recent research employing experimental designs has

demonstrated that other lending models, including

noncollateralized loans to individuals, can achieve

similarly impressive results (Giné and Karlan 2012).

This has shifted focus away from the joint liabil-

ity mechanism to more general contractual structures

common in microﬁnance (Giné et al. 2010).

Recent research shows that future access to cap-

ital is the central incentive for timely repayment,

which can be reinforced through different contractual

structures, including (a) lending progressively (i.e.,

loans increase in size with timely prior payment) to

place more value on future loans than on a potential

default, (b) favoring borrowers (e.g., women’s groups)

who generally follow more conservative investment

strategies, (c) making small yet frequent payments

(e.g., weekly), or (d) engaging in intense supervi-

sion of clients (Armendáriz de Aghion and Morduch

2000,2004;Anthony 2005;Field and Pande 2008;

Giné et al. 2010).

Considerably less attention has been devoted to the

organizations and the agents that create and enforce

MFI contracts (Jain and Moore 2003). This neglect

is surprising for several reasons. First, microﬁnance

clients are mostly poor and geographically dispersed,

which means that loan ofﬁcers are often the sole point

of contact between a client and the MFI. Whereas

the central MFI determines lending policies, loan con-

tracts, and collection procedures, these are always

enacted and primarily enforced by loan ofﬁcers oper-

ating from a network of small branches. It is loan

ofﬁcers who ﬁnd and evaluate clients, gather credit

information, perform a credit analysis, produce a

credit recommendation, and provide all the neces-

sary information for the disbursement of the loan.

They also supervise loan repayment, ensure collec-

tion in cases of missed payments, renew and increase

credit lines when a loan reaches maturity, and “sell”

other products such as life insurance to good clients.

As a result, even though clients have a contractual

agreement with the MFI, they establish and experi-

ence it through their relationship with loan ofﬁcers.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Second, client needs vary considerably across indi-

viduals and time, making ex ante contracts necessar-

ily incomplete. More broadly, microﬁnance constitutes

an extreme version of small business ﬁnance, which

has been shown to be especially sensitive to relational

contracts and the “soft” information that can only

travel through personal—rather than contractual—

ties (e.g., Petersen and Rajan 1994,Berger et al.

2001). It also shares important characteristics with

street-level bureaucracies, where agents must exercise

unusual levels of discretion in the enactment of rules

(Lipsky 1980,Canales 2011,Coslovsky 2011,Piore

2011). For these reasons, we can expect that borrower–

loan ofﬁcer relational considerations will interact

with formal contractual structures to affect loan out-

comes. In addition, the high frequency of microcre-

dit results in a large number of contractual exchanges

between an MFI and its diverse client population

(Giné et al. 2010). Microﬁnance therefore presents

a rich setting to test and separate speciﬁc claims

and mechanisms concerning formal and relational

contracts.

1.2. Predictions Concerning Contracts,

Loan Ofﬁcer Change, and Relational Styles

Contracts are foundational components of economic

and social life (Durkheim 1997, p. 155; Weber 1978,

Chap. VIII; Fudenberg and Tirole1990). Contracts spec-

ify rights and responsibilities between individuals—

or between individuals and institutions—in various

exchanges. From the most mundane of matters such

as consumer purchases to the most intimate such as

marriage, contracts reﬂect, shape, coordinate, and cir-

cumscribe expectations and action.

In a spot market with perfect information, contracts

would not be essential because there would be no

need to account for contingencies or facilitate coor-

dination (Hermalin et al. 2007). In the presence of

information asymmetries and the risk of contingen-

cies, however, contracts can stipulate the price, quan-

tity, and timing of repayment, as well as penalties and

remedial rights if the contract is breached, thus facil-

itating exchange.

Formal contracts have limitations. First, they are

incomplete, as the myriad contingencies that may

arise cannot be accounted for ex ante (Hart and Moore

1999). Second, they can be costly in time and money

to craft and enforce. Relational contracts solve some

of these limitations. They are less rigid than for-

mal contracts as they rely on the trust that devel-

ops between parties who have an interest in repeated

exchange (Macaulay 1963,Macneil 1978,McMillan

and Woodruff 1999,Baker et al. 2002).

Some have argued that formal contracts and the

trust required for relational contracts are substitutes

(Zucker 1986,Guseva and Rona-Tas 2001). Others

have argued that formal contracts stunt the devel-

opment of trust (Malhotra and Murnighan 2002).

Research also suggests that the two can be comple-

ments (e.g., Poppo and Zenger 2002) and are often

used to reinforce each other (Baker et al. 1994). A case

in point is the imagery of community bankers who

rely on relational lending practices that incorporate

“soft” and “hard” information in decision making, as

well as formal contracting mechanisms (e.g., Sharpe

1990,Berger and Udell 1995).

Relational contracts can be between organizations

or people and organizations. For example, IBM

once offered the promise of “lifetime employment”

(Baker et al. 1994) that was not formally stipu-

lated in employment contracts but was understood

by corporate actors and applicants who enacted

and enforced these contracts. More generally, orga-

nizational blueprints entail a host of implicit and

explicit agreements with employees concerning the

“employment deal,” such as how their efforts will

be coordinated, controlled, and compensated (Baron

et al. 2001). When blueprints are changed, employee

turnover often ensues as implicit agreements are

rescinded.

Change is ubiquitous in organizations as employ-

ees come and go, strategies and blueprints change,

and internal soft expectations of performance adjust to

align with the demands of changing market environ-

ments. When change alters or severs the ties that have

sustained a relational contract, it follows that the value

it creates can be compromised, often leading to the

loss of an exchange relationship (e.g., Seabright et al.

1992,Hannan et al. 1996,Baron et al. 2001,Broschak

2004,Biong and Ulvnes 2011).2This follows because

relational value is created by individuals who play

dual roles, both as individuals and as representatives

of their organizations. When a loan ofﬁcer changes an

aspect of the relationship, for example, the borrower

determines whether to confer ownership of the rela-

tional value to the loan ofﬁcer or the organization the

loan ofﬁcer represents (Sorenson and Rogan 2014).

As noted above, microﬁnance loans are simple in

form: they have ﬁxed terms, rates, and a straight-

forward amortization schedule summarized in a sim-

ple legal document. Timely payment should thus

not vary as a function of the loan ofﬁcer because

the contract is signed between the borrower and the

MFI.3More broadly, credit methodologies typically

2The MFI (and industry) we study experiences high rates of

employee turnover (Janik 2012). Consequently, it seeks to limit the

depth of ofﬁcers’ ties to clients because it fears that if they leave

they will take good clients with them. Random loan ofﬁcer rota-

tions are employed to reduce this possibility. But, as a consequence,

MFIs suffer increased delinquencies, as we show here.

3Payment is made directly, electronically, to the MFI. Consequently,

there is no technical reason changing loan ofﬁcers should have an

impact on timely payment.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

assess the credit worthiness of a borrower. Clients

can be classiﬁed as “good credit risks” because client

types are assumed to be stable. To keep track of

clients, facilitate loan ofﬁcer work, and ease transi-

tions in cases of turnover, MFIs—including the ones

we study—have developed sophisticated software

platforms that standardize lending decisions, keep

track of loan activity, and enable loan ofﬁcers to col-

lect and record rich information about each client and

document every visit or interaction. There is thus no

economic, informational, or legal reason why clients,

who were evaluated with the same credit methodol-

ogy, should differ in their repayment behavior when

interacting with loan ofﬁcers with different relational

styles. Furthermore, given that lack of payment will

likely lead to the loss of the credit line, the signiﬁcant

incentive for repayment provided by future access to

credit is in no way dependent on a loan ofﬁcer’s rela-

tional style. Finally, loan ofﬁcers are in a position of

power as a function of the client’s dependency on

them (Emerson 1962). Formal mechanisms and power

dynamics should thus strongly incent timely payment

irrespective of the relational styles employed by loan

ofﬁcers.

Yet our ethnographic work revealed the negative

impact of change on borrowers. As noted above,

although borrowers know their loan is provided by

the MFI, they often perceive that their commitment

is to the loan ofﬁcer they have worked with. Thus, a

change in the loan ofﬁcer can be experienced by the

client as a change in the relationship with the organi-

zation, which reveals the presence of relational con-

siderations alongside their formal contract.

1.3. Individually Consistent Relational Styles

(Within Individual)

The ways in which social actors interact can be ab-

stracted and classiﬁed based on the roles each occu-

pies (Berger and Luckmann 1967, pp. 72–77). This

means that although there is heterogeneity in social

interaction, there is also much that can be assumed

based on homogeneity within roles. We argue that

similar dynamics operate with respect to relations.

Social cognitive psychologists have focused on spe-

ciﬁc layers of this process and refer to the oper-

ating mechanism as relational schemata (Baldwin

1992,Sanchez-Burks et al. 2000). Relational schemata

include an interactional model of “ego” and “alter.”

Each actor has expectations of self and other, and

both employ relational scripts that form a model of

a typical interaction (Smith 1984). Relational scripts

are often conceptualized as “if-then” statements that

structure behavior. Hence, each interaction, even

those with complete strangers, starts with a baseline

of understanding. We focus on the aggregation of

these schemata into systematically reoccurring rela-

tional styles used by actors across interactions.

In ﬁeld observations of loan ofﬁcers, we found that

certain loan ofﬁcers interpret and enact policies ﬂexi-

bly while others adhere strictly to them. These obser-

vations led to an inductively developed typology of

loan ofﬁcer relational styles. Table 1presents their

underlying interpretive frames. One category of loan

ofﬁcers tends to adhere strictly to the rules. They rely

heavily on standardized models to assess borrowers,

customer relationship management systems to deﬁne

client management tasks and track client information,

and handheld technological devices to help automate

decision making in line with organizational policies

and procedures. In turn, this underlying philosophy

informs and is enacted in their relationships with

clients. We refer to them as “letter of the law” (LL)

ofﬁcers. Consider a typical LL ofﬁcer’s description of

his job:

My job is to recruit the new loan groups, train them on

the methodology, do the credit analysis, and make sure

that the whole process runs smoothly. It basically con-

sists of applying the methodology strictly and making

sure that the groups adhere closely to the policies.

LL loan ofﬁcers adhere strictly to rules not nec-

essarily because they believe there is no latitude in

their client relationships. Rather, they assume that

the rules are designed intelligently to maximize efﬁ-

ciency. The interpersonal distance maintained by LL

ofﬁcers is intended to maintain objectivity in rela-

tions with limited information, not necessarily out

of contempt for clients. Ofﬁcers in the second rela-

tional category, “spirit of the law” (SL), interpret pro-

visions more expansively and ﬂexibly; they also tend

to develop deeper relationships with their clients as

they learn about their lives, needs, and concerns holis-

tically. This is not necessarily out of altruism. SL ofﬁ-

cers believe developing multiplex relationships with

clients enables them to perform their jobs more effec-

tively (May and Winter 2000):

Policies are good, they are useful, but you also need

to give them a personal touch. 000 It is much better

to spend that time learning about the client, about

her relationships, and selling her on opportunities.000

That’s what helping my clients is about. (SL ofﬁcer)

Finally, some loan ofﬁcers exhibited relational styles

that are neither purely LL nor SL, but display a mix

of both. These loan ofﬁcers were labeled as “mixed,”

as they blend elements of each style with every client,

rather than different styles with different clients.4

4We interviewed managers and loan ofﬁcers after the typology was

developed. A few notes are worth making. First, they all believe

that loan ofﬁcer styles are stable. This is because styles emerge from

an underlying philosophy about the nature of organizational rules

and the role of the MFI vis-à-vis its clients. Furthermore, clients

adapt and react to styles, generating self-reinforcing dynamics.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Table 1 Interpretive Frames Associated with Different Relational Styles

Why and how loan

ofﬁcers interact as

they do with clients Spirit of the Law (SL) Letter of the Law (LL)

View on the nature of

microﬁnance and

borrowers

—The best thing for the organization is to help its clients. The

organization should always adapt to meet client needs.

“To the clients, we are 6MFI7. And to 6the MFI7, we are the

clients.000 We have to make sure that the company is doing

the right thing, that it is truly helping our clients, and that our

clients are being heard.”

—The best thing for the organization is to stick to the business

model that it knows and that works. The organization should

select clients who ﬁt the model well.

“At the end of the day, we are a business. We are not a welfare

organization, even if our work helps people.”

View on organizational

rules and related

client interaction

—Rules are tools, to be used as needed.

—Rules are made by people who have less information than is

available on the ﬁeld.

“Policies are good, they are useful, but you also need to give

them a personal touch. 000 Managers don’t see what I see.

The policies don’t see what I see.”

—Rules are instructions that specify what can and cannot be

done.

—Rules are made by people who have more information than is

available on the ﬁeld.

“I explained things to them carefully; I went over the contract

with them and they agreed. I told them of the responsibilities

and the implications and they agreed. I went through different

scenarios of good things and bad things that could happen

and what the contract speciﬁed for good behavior and for bad

behavior, and they agreed.”

Contracts,

information, and

clients

—Our rules and contracts are incomplete in unforeseen ways.

The more information I can have about my client—including

personal information—the better.

—The more information I can reveal to my client about me, how

we do things, and how the system works, the better.

“When you visit a prospect’s store, you could just begin asking

questions following the policies.000 It is much better to spend

that time learning about the client, about her relationships,

and selling her on opportunities, such as the purchase of a

ham slicer and how productive the loan would be.”

—The only information that is relevant is that which is speciﬁed

and codiﬁed in our contracts and our systems. Collecting

other types of information is a waste of our time.

—Our clients need to understand how our contracts work and

what their formal commitments are.

“There is a reason why we have these tools. The most efﬁcient

thing to do is to just ask the questions that the system needs

in that order. You input information, the tool gives you an

answer, and you’re done.”

How loan ofﬁcers

handle problems

in the client

relationship

—We work with the client to understand the situation and jointly

ﬁnd a solution; the contract is a guideline.

“Whenever my poorer clients tell me they cannot make a

payment because something bad happened to them, I have a

policy of always trusting them. So I sit down with them,

understand the problem and we come up with a solution,

even if that means ﬁghting 6the MFI7to change the terms of

the loan or do a formal restructuring.000 Sure, some of my

clients end up not being morally solvent, but I can tell you

that of every 10 clients I have helped, 9 have made it and 8

have become long-term clients.”

—The client knows what the rules are; we enforce the contract

as speciﬁed.

“I did not design these rules, and I did not force my clients to

sign these contracts 000what’s the surprise? I am only doing

what I told you 000keeping my side of the contract that you

and I signed.” “Once I make an exception for the ﬁrst woman,

what do I do when her cousin calls me and wants the same?

000No, no, no. We have an agreement. I don’t come in and tell

you why I did not deposit your loan amount after we signed a

contract. So, if you don’t repay the company, my job is to

come in and make sure that you do.”

Notes.N=711: SL =235, LL =233, and 243 =mixed. “Mixed” loan agents blend various elements of SL and LL agents, vacillating between the two. The

em dash denotes a summary of approach across observations. Italicized text represents a representative quote.

Substantial research has documented the sociocog-

nitive processes that operate in dyadic interactions

when there is ambiguity in action or expected reaction.

Parties to such exchanges draw on prior experiences

and underlying expectations to inform current interac-

tions. Theory and research argue that individual social

actors are evaluated based on their salient characteris-

tics and others’ experience with those characteristics.

For example, employers extrapolate the potential of

prospective employees based on their speciﬁc gender

or racial characteristics (see, e.g., Becker 1959,Arrow

Finally, loan ofﬁcers manage, on average, 250 clients. It is cogni-

tively taxing—some of them argue impossible—to follow different

styles with different clients.

1972,Bielby and Baron 1986,Fernandez and Green-

berg 2013).

A similar dynamic is evident in other social interac-

tions where actors have no prior relationships, which

leads them to extrapolate from typologies of prior

experience. Police ofﬁcers, for example, use different

scripts depending on which type of citizen they per-

ceive they are dealing with, so they develop heuristics

to determine whether they are likely dealing with a

citizen, a criminal, or an “asshole” (Van Maanen 1978).

In microﬁnance, loan ofﬁcers similarly try to estab-

lish role and relational expectations about their clients

when they evaluate them to create a proﬁle. Clients

might be classiﬁed as high-risk or low-risk based

on observable factors such as industry, proposed

use of funds, and credit scores, as well as on soft

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

information. Once classiﬁed, borrowers’ subsequent

actions are often interpreted with reference to these

initial classiﬁcations.

For a borrower who most likely has little experience

with ﬁnancial products, it is also important to under-

stand the expectations that her loan ofﬁcer has of her

because compliance with their relational contract will

deﬁne her ability to succeed as a client and, therefore,

access future loans from the MFI. Loan ofﬁcers must

therefore clarify task knowledge to clients on the techni-

calities of a loan and provide them with clear relational

knowledge (Gibbons and Henderson 2011) about how

those technicalities will be enacted. Will her loan ofﬁ-

cer establish and expect an arm’s-length exchange? Or

will the exchange rely on a closer social relationship

where the expectation is one of joint problem solv-

ing, transmission of soft information, and a contextual

interpretation of contractual terms?

By relational styles we refer to the speciﬁc manner in

which social actors in given roles interact and relate

to other social actors in given roles. Focus here is not

on the strength of the relationship (Granovetter 1973),

nor of the functional or other characteristics of the role

incumbent or successor (e.g., Burton and Beckman

2007). Rather, it is on the speciﬁc ways in which actors

interact, and how this interaction informs future inter-

actions. For example, a loan ofﬁcer, by virtue of prior

experience, has a cognitive blueprint she enacts via

scripts when interacting and responding to different

client actions. This blueprint is based on her interpre-

tation and understanding of organizational policies,

incentives, and personal feelings about how to relate

to clients. Loan ofﬁcers can signal these expectations

through a clear and consistent relational style, and

clients can adjust their behavior accordingly. One loan

ofﬁcer explained,

They care about the loan and about the money, but

they worry more about whether things are going to

work out for longer, you know? They want the oppor-

tunity, not just the money. [000] So, you have to teach

them. You see them begin to manage the loan, and you

have to be clear with them, sometimes repeat the same

thing several times. Then they start becoming more

astute when they manage their money, and then you

see them loosen up; you see them trust you because

you do what you told them you would do. They do

what you told them to do, and then there are no tricks;

you increase the amount when they ﬁnish the ﬁrst

loan and things happen as you described them, so they

trust you.

On the other hand, if a loan ofﬁcer does not signal

clear expectations, or sometimes employs relational

scripts associated with a letter of the law style and

other times a spirit of the law style, borrowers will

have a difﬁcult time understanding soft expectations

and respond accordingly (i.e., the if-then link becomes

unclear). A borrower accustomed to a spirit of the law

approach to contractual enforcement may thus come

to expect that if he confesses experiencing some famil-

ial challenges that undermined his ability to pay on

time his SL loan ofﬁcer will then express understand-

ing while devising a plan to get the borrower back

on track. This expectation would, however, be under-

mined if the loan ofﬁcer responds by indicating her

intent to strictly enforce contractual terms. Moreover,

this inconsistency should lead to ambiguity, misun-

derstanding, or even mistrust about what the loan

ofﬁcer expects of the borrower in future interactions.

This leads us to expect the following.

Hypothesis 1 (H1). A borrower is less likely to miss

a payment contrary to contractual terms when her loan

ofﬁcer has an individually consistent relational style.

1.4. Along Consistent (Between Individual)

Relational Styles

So far, we have speciﬁed how (within-ofﬁcer) consis-

tency helps clarify expectations and thus provides a

solid basis for relational contracts. Consistency clari-

ﬁes expectations because the elements that comprise

the actor and action are logically interconnected. The

relational scripts employed by a letter of the law loan

ofﬁcer are mutually reinforcing manifestations of a

rational philosophy of action with certain premises

about behavior. For example, one LL ofﬁcer described

the relational script he employs with clients:

I explained things to them carefully, I went over the

contract with them and they agreed. I told them of the

responsibilities and the implications and they agreed.

I went through different scenarios of good things and

bad things that could happen and what the contract

speciﬁed for good behavior and for bad behavior and

they agreed.

Consistency is concerned with the regularity of

style or action irrespective of its content. It clariﬁes

expectations by providing a within-actor baseline for

prediction. For example, prior actions by counterpar-

ties who follow a particular relational style lead to

predictions about their likely future relational styles,

as evidenced by models of relational embeddedness

(e.g., Uzzi 1997,Broschak 2004,Broschak and Block

2014). A broken tie is disruptive precisely because it

removes the dyadic history that sustains expectations.

There is nothing in this deﬁnition, however, to

restrict consistency to repeated interaction with the

same social actor. Sociocognitive understanding can

also arise from interaction with successive actors

who have similar relational styles. We label this

“along consistent relational styles” (i.e., consistency

in the individual styles employed by successive indi-

viduals), building on Max Weber’s (1962) concept

of uniformity concerning interaction with different

individuals.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

An implication of a focus on relational styles, there-

fore, is that trust and understanding can emerge

through dealings with different actors who exhibit a

similar way of interacting in speciﬁc roles and situ-

ations. This intuition underlies economic and socio-

logical models of employer screening that incorporate

consideration of categorical characteristics (e.g., race,

gender), as noted above. Another example is the com-

parative advantage that an organization gains from

screening for and then inculcating and rewarding sys-

tematic relational styles that reﬂect its identity in its

employees. Greetings by employees of Soup Kitchen

International (as demonstrated by the “Soup Nazi” in

the Seinfeld sitcom), for example, are of a deﬁnable

type that attracts customers in addition to the food.

In a different business context, Southwest and Sin-

gapore Airlines have different value propositions,

brand identities, and organizational cultures. The for-

mer stresses low-cost, no-frills air transportation that

is fun and requires customer involvement to achieve

cost advantages (e.g., Heskett and Hallowell 1993).

The latter positions itself as a premium product and

service. Although the functional objective of each of

these businesses is to transport people, how they do

so from a customer-experience perspective is drasti-

cally different. Consequently, so are their value propo-

sitions. Yet the sum of their internal processes and

external identity are tightly aligned in consistent yet

different ways, and both are praised for their excep-

tional yet distinct experiences.

Indeed, a ﬂight attendant who previously worked

at Singapore Airlines and switched to Southwest

would have a confusing experience when ﬁrst inter-

acting with Southwest customers and employees, and

vice versa. Moreover, in dealing with various employ-

ees of each company, a deﬁnable, consistent relational

style becomes evident that employees and customers

may have a preference or distaste for. Problems arise

when “what is said” deviates from “what is done.”

The relational styles evident in these organizations

are no accident. Organizations spend considerable

resources selecting and teaching employees how they

should interact with clients (Van Maanen 1973,1978,

1991). In our research context, trainees spend several

months shadowing experienced loan ofﬁcers, observ-

ing and learning how they manage clients in the ﬁeld.

With experience, loan ofﬁcers become keenly aware of

the importance of providing a consistent client expe-

rience, as described by an ofﬁcer:

Once clients become used to working in one way, if

you change things, many of them get confused and

some even go to another MFI. 000I know I am a good

loan ofﬁcer, but my style is different, and I learned that

the hard way, losing clients in a previous rotation.

Consistency in the relational styles used by consec-

utive loan ofﬁcers in our setting should thus facilitate

borrower understanding even in the face of a bro-

ken interpersonal tie because the schemata and scripts

employed by different loan ofﬁcers are internally con-

sistent and may also be similar, and thus the client

has a sounder basis for understanding expectations

even without any prior experience with a speciﬁc loan

ofﬁcer. Conversely, inconsistency in relational styles

should increase interpretive difﬁculty in soft expec-

tations or the enforcement of hard stipulations, thus

increasing ambiguity. One client who experienced a

loan ofﬁcer change explained,

[Turning to his wife] What was the name of the other

woman? Yes, she was nice. I really liked how she

worked. She came in, it was all business: Do you want

to renew? How much? Any problems? Any referrals?

And she was gone. That worked for me. But this new

guy, man, he is so chatty. Always asking me about

other stuff, and then telling me how to run my busi-

ness. I just want my loan to run smoothly, as it did

with [the previous ofﬁcer]. So yeah, the change was

not great for me. (Borrower)

Inconsistency across loan ofﬁcers should increase

the likelihood of delinquency as it becomes more dif-

ﬁcult to predict how a subsequent loan ofﬁcer will

enforce contractual terms. This follows because not

only must the borrower become acquainted with a

new loan ofﬁcer she does not know but she must

also learn how to interpret and thus respond to the

new loan ofﬁcer’s requests and demands. The latter

is more difﬁcult when the borrower’s previous ofﬁcer

and new loan ofﬁcer have distinct relational styles,

which leads to distinct expectations and responses to

contingencies. We should thus observe the following.

Hypothesis 2 (H2). A borrower is less likely to miss

a payment contrary to contractual terms when there is a

change in her loan ofﬁcer and the subsequent loan ofﬁcer

has a relational style consistent with the previous loan ofﬁ-

cer’s style.

From the discussion above, we can infer that the

disruption caused by a broken tie that brings inconsis-

tency to a borrower’s experience will be even larger

when this disruption also brings with it an internally

inconsistent new set of expectations. Such disruption

further compounds and increases the level of ambigu-

ity about expectations and likely reactions as a client

must contend with change on two margins: ﬁrst, the

new loan ofﬁcer does not engage interpersonally in

the way her predecessor did in response to similar sit-

uations; and second, the loan ofﬁcer herself engages

with the client differently across situations that seem

similar. By contrast, this is less acute or moderated

when a change in loan ofﬁcers and styles does not

also introduce internal inconsistency. We should thus

expect the following.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Hypothesis 3 (H3). A borrower is less likely to miss

a payment contrary to contractual terms when there is a

change in her loan ofﬁcer, the subsequent loan ofﬁcer has

a relational style inconsistent with the previous loan ofﬁ-

cer’s style, but the subsequent loan ofﬁcer’s style is also

individually consistent.

It is worth noting that relational styles, and social

action more generally, can be consistently inconsis-

tent,5which implies a set of scripts not applied in a

logically interconnected manner. We argue above that

individually (or intra-loan ofﬁcer) consistent styles are

preferable, as is consistency across different loan ofﬁ-

cers (i.e., inter-loan ofﬁcers). These two propositions

lead to a potential tension when there is a movement

from individually inconsistent to individually consis-

tent relational styles, which also implies along-actor

inconsistency. One can imagine two plausible predic-

tions concerning which basis of predictability is most

useful, with each favoring the two variants of consis-

tency discussed above. Absent theory, we pose this as

an empirical question to be explored with the data at

hand rather than as a formal hypothesis; to wit: Is a

borrower comparatively less (more) likely to miss a

payment contrary to contractual terms when there is

a change in her loan ofﬁcer who has an inconsistent

style and the subsequent loan ofﬁcer has an individ-

ually consistent relational style?

2. Setting and Analytical Strategy

2.1. Data

The quantitative data used in this study are drawn

from a proprietary loan-level database maintained by

one large MFI in Mexico for the period 2004–2008.

The MFI is a well-established and regarded market

leader that provides mostly individual loans in urban

areas, primarily Mexico City. The database includes

extensive, standardized measures for 450,000 loans

administered by more than 700 loan ofﬁcers, which

reﬂects all the loans administered by the MFI in the

study time frame. On average, each loan ofﬁcer han-

dles approximately 250 borrowers at a given time,

affording a sufﬁcient number of transitions between

varying combinations of loan ofﬁcer relational styles

for robust estimation. In terms of relevant informa-

tion, the measures include loan size, size of scheduled

payments, interest rate, term, client payment history,

client gender, whether the loan was administered as

part of a group, and previous delinquency. We focus

on loans as the level of analysis because each has spe-

ciﬁc features (e.g., interest rate, term, size) that vary

5On related intuition concerning predictably irrational behavior,

see Ariely (2008).

within borrowers and can have a bearing on delin-

quency.6Moreover, ultimately, loan ofﬁcers must jus-

tify speciﬁc loans in branch credit meetings.

Two features of the study setting are worth high-

lighting: First, loan ofﬁcers are randomly assigned

to branches upon entry to the MFI and, once there,

are assigned a geographic area of coverage. Second,

the company has a policy of randomly rotating loan

ofﬁcers across branches. For example, the MFI pur-

posefully staffs new branches with existing, randomly

selected loan ofﬁcers from the pool of hires and then

rotates additional ofﬁcers around vacancies. The same

applies when a loan ofﬁcer leaves the MFI, which

occurs quite frequently. This is a common policy in

MFIs employed to reduce the risk of corruption and

collusion between loan ofﬁcers and clients, as well as

the “capture” of clients by loan ofﬁcers who can take

them to a competitor if they leave the ﬁrm. When

rotations happen, the loan ofﬁcer is assigned an exist-

ing portfolio at her new branch (see Hertzberg et al.

2010 and Fishman et al. 2011 on the logic of rotation

practices). Her original portfolio is then assigned to

a new loan ofﬁcer or split between the ofﬁcers who

remain in the branch.

Our basis of identiﬁcation is thus the fact that re-

gardless of why a tie was broken between a client

and initial loan ofﬁcer, the subsequent loan ofﬁcer is

assigned at random. In turn, she may or may not have

a relational style that is consistent with the prior loan

ofﬁcer, and it is this (in)consistency that is of inter-

est here. This provides random variation with respect

to the (in)consistency in the previous and subsequent

loan ofﬁcers’ relational styles, which is orthogonal

to client and loan ofﬁcer characteristics or the basis

for the (unobservable) cause of the originating loan

ofﬁcer’s vacancy, that affords the analytical leverage

required to test Hypotheses 2and 3concerning incon-

sistency in inter-loan ofﬁcer relational styles. This

exogeneity cannot, however, be substantiated with

respect to our ﬁrst prediction about inconsistency in a

speciﬁc (i.e., intra) loan ofﬁcer’s relational style, which

nonetheless relies on tests including careful inclusion

of theoretically informed covariates.

This paper is part of a larger mixed-data research

program considering processes and outcomes in mi-

croﬁnance. The qualitative component of the project

includes ethnographic evidence and interviews with

MFI managers and loan ofﬁcers (N=76) at differ-

ent levels of experience and institutional authority,

as well as a random subset of their best and worst

clients (N=53). We conducted about 400 hours of

interviews and signiﬁcant ethnographic observation

6Treating borrowers as the unit of analysis would therefore require

ignoring loan and time-varying contextual information or aggregat-

ing data in some fashion that would obscure important differences.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

at three different MFIs in Mexico, generating approx-

imately 1,200 pages of notes that were transcribed by

one of the authors. These qualitative data are used

here to frame hypotheses and to inform the develop-

ment, understanding, and interpretation of key con-

structs, ﬁndings, and implications. The qualitative

data have been discussed in other studies (Canales

2011,2014). Therefore, we limit presentation and dis-

cussion of them here to conserve space and focus

attention on the novel contribution of this paper—the

implications of within-loan-ofﬁcer and between-loan-

ofﬁcer consistency on delinquency.

2.2. Measures

2.2.1. Outcome Measures. We calculate three

dummy-coded outcome measures that evidence

breach of contract: the ﬁrst denotes that the bor-

rower missed one payment or more in breach of

contract during the life of the loan (mean =0.333,

SD =0.471). The second denotes that the bor-

rower missed two or more consecutive payments

(mean =0.168, SD =0.374). The third denotes that the

borrower missed three or more consecutive payments

(mean =0.115, SD =0.319). As evidenced by these

summary statistics, a third of the borrowers miss a

payment.7It is worth noting that of those who miss

a payment, nearly half miss a second payment. And

contingent on missing a second payment, two-thirds

miss a third payment or more. To conserve space,

we discuss the ﬁrst measure in depth while making

reference to the other outcome measures in passing.

Complete results for these models are available in the

online appendix (available as supplemental material

at http://dx.doi.org/10.1287/mnsc.2015.2167).

These measures are well suited to assess a breach

of contract and the theoretical predictions outlined

above because the timely repayment of a loan is the

most important responsibility of the borrower. In fact,

repayment rates are the ﬁrst and most common met-

ric used by MFIs to evaluate loan ofﬁcers, branches,

and the overall health of their lending portfolio. More

broadly, the timely repayment of a loan is the primary

measure of compliance in any credit relationship in

Mexico or the United States.

2.2.2. Predictors. We calculate two sets of vari-

ables to classify individual loan ofﬁcer consistency

in relational styles. The ﬁrst set includes spirit of

the law (SL), letter of the law (LL), and mixed (M)

7This number may seem high compared with microﬁnance best

practices, which usually document delinquencies below 5%. Publi-

cized numbers usually focus on 30-day or 60-day delinquency rates,

whereas we focus on the dynamics of enforcement by observing

trajectories of delinquency from the ﬁrst missed payment. Overall

rates of delinquency and default mirror standard best practices.

relational style dummy variables to capture individ-

ual (within-actor) consistency of relational styles. SL

and LL styles are regarded as consistent and distinct

relational styles guided by principles that subsume

scripts that “hang together” in a logical fashion. By

contrast, loan ofﬁcers exhibiting a mixed style engage

in practices that incorporate elements of both SL and

LL styles—practices that often do not hang together

in a logically complementary manner. For example,

a mixed loan ofﬁcer might ask similar questions and

signal a similar personal relationship with a client as

an SL ofﬁcer but may be unresponsive if the client

faces a problem that makes it difﬁcult to pay, whereas

an SL ofﬁcer would almost invariably engage in joint

problem solving. By contrast, an LL ofﬁcer will typi-

cally use formal contractual mechanisms (asset conﬁs-

cations, the presence of company collectors) to compel

repayment.

To code loan ofﬁcers according to their relational

styles, three regional managers who supervise and

know all the loan ofﬁcers were independently shown

Table 1and, using the full roster of loan ofﬁcers, were

asked to code each loan ofﬁcer as either SL, LL, or

mixed. We emphasized that the categories referred to

their relational styles and not to their performance.

It is worth mentioning that managers thought the

typology was descriptive, intuitive, and a fair repre-

sentation of loan ofﬁcers. Managers also coded ofﬁ-

cers quickly, which both reinforces the validity of the

typology and reveals the depth with which managers

know their staff. Interrater reliability was just below

80%. At the same time, no manager had ever coded

loan ofﬁcers this way, so managers did not have intu-

ition about which style would be “better.” In fact,

as documented in other work (Canales 2014), when

asked at the end of the process which style they

anticipated would perform best, every manager had

a different theory. There was no instance where one

manager coded an ofﬁcer as SL while another coded

her as LL. The only discrepancies were between indi-

vidually consistent styles (LL or SL) and mixed. These

discrepancies were treated as mixed. The reason for

doing so is straightforward: the inability of managers

to agree on how to classify a loan ofﬁcer provides

prima facia evidence that the loan ofﬁcer’s style is

individually inconsistent.8

To ensure this classiﬁcation captures real differ-

ences in loan ofﬁcer relational styles rather than loan

ofﬁcer experience, human capital, or other charac-

teristics that may be correlated with managers’ per-

ceptions of loan ofﬁcers’ relational styles as well as

their loan portfolio delinquency rates, we performed

8See the online appendix for more information concerning con-

struct development and validity checks, as well as for ethnographic

information on typology development.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

a number of tests reﬂected in Table A1 of the online

appendix, in which we also discuss construct validity

in more depth. First, we calculated a direct measure

of loan ofﬁcer experience (tenure) and then compared

it across styles reasoning that experience may be cor-

related with managers’ ability to accurately classify

loan ofﬁcers. Second, we looked at the trajectory of

loan ofﬁcers within the MFI, assessing whether loan

ofﬁcers with different styles were rotated at different

rates (total branch rotations), deserted the ﬁrm in dif-

ferent proportions or rates (turnover in percentage)

and turnover in days), or had a different initial train-

ing experience (time in the ﬁrst branch).

We also looked at other loan ofﬁcer observable

characteristics, including educational attainment, age,

marital status, and gender. We calculated and com-

pared average loan values and interest rates across

styles. We reason that if managers differ in their

knowledge of loan ofﬁcers’ experience and task-

speciﬁc human capital, thus leading to misclassiﬁ-

cation, managers should be more restrictive in the

amount of the MFI’s resources they allocate to those

loan ofﬁcers for the same reason. The results of

these tests conﬁrm that managers classify loan ofﬁcers

according to relational style and not based on a lack of

experience with, or knowledge of, the ofﬁcers. Finally,

we directly compared loan ofﬁcer performance by

relational styles to determine whether the mixed cat-

egory is a proxy for underperformance. Results are

inconsistent with such a supposition. On average,

mixed loan ofﬁcers received average bonuses based

on their loan portfolio performance that did not differ

statistically from LL ofﬁcers, for example. Moreover,

the loan ofﬁcer with the single highest bonus was cat-

egorized as exhibiting a mixed style, and numerically

speaking, more mixed ofﬁcers (126) performed in the

75%–25% interval than either SL (108) or LL (117)

ofﬁcers.

The second set of measures assesses the impact

of broken ties on loan repayment. A variable codi-

ﬁes cases when a client was transferred to a differ-

ent loan ofﬁcer (changed ofﬁcer). We then include

interactions capturing between-actor consistency in,

or the transition from, one loan ofﬁcer relational

style at time tto the same or a different style at

time t+1. We create summary and ﬁne-grained mea-

sures. The summary measures are dummy variables

denoting (1) between-actor consistency in individu-

ally consistent relational styles (SLt→SLt+1or LLt→

LLt+1), which means that although the client has a

new loan ofﬁcer, she employs the same relational

style as her predecessor (“along consistent” relational

styles); (2) individually consistent, successively incon-

sistent styles (SLt→LLt+1or LLt→SLt+1) (across

consistent or “individually consistent to individually

consistent”); (3) along individually inconsistent styles

(Mt→Mt+1) (“individually inconsistent to individu-

ally inconsistent”); and (4) moving from a within-

actor individually inconsistent style to an individually

consistent style (Mt→SLt+1or Mt→LLt+1) (“incon-

sistent to consistent”), which denotes that the original

loan ofﬁcer employed an inconsistent relational style

whereas the subsequent one employed a consistent

one—either letter or spirit of the law.

The ﬁne-grained transitions include spirit of the law

to spirit of the law (SLt→SLt+15, letter of the law to

letter of the law (LLt→LLt+15, mixed to spirit of the

law (Mt→SLt+15, mixed to letter of the law (Mt→

LLt+15, letter of the law to mixed (LLt→Mt+15, letter

of the law to spirit of the law (LLt→SLt+15, spirit of

the law to letter of the law (SLt→LLt+15, and spirit of

the law to mixed (SLt→Mt+15, with the omitted cate-

gory in most speciﬁcations being changes from mixed

to mixed. These measures provide a means of assess-

ing how consistency in style across different loan ofﬁ-

cers over time has a bearing on the probability of

delinquency (see Baron et al. 2001 and Burton and

Beckman 2007 for similar coding strategies).

2.2.3. Controls. We include three classes of con-

trols, which are described and tabulated in Table 2,

to help rule out alternative explanations and some

endogeneity concerns. One class includes measures of

the characteristics of the loan that may have a bear-

ing on the recipient’s ability and willingness to repay

in a timely manner, which helps alleviate concerns

of reverse causality pertaining to client performance

and loan ofﬁcer turnover. These include the size of

the loan in thousands of pesos (mean =ln4809255,

SD =8.561); gender (female =00624, SD =0.484); days

between scheduled loan payments (mean =15.545,

SD =14.423, range =7–86) (Field and Pande 2008);

whether the loan has been issued to a group, which

may increase social pressure to adhere to the pro-

visions of the loan (mean =0.15, SD =0.357) (e.g.,

Stiglitz 1990,Armendáriz de Aghion 1999,Sanyal

2009; but see Giné and Karlan 2012); interest rate

charged (mean =80.122, SD =6.476, range 58–96);

and a count of the total number of previous pay-

ments the borrower missed (mean =1.083, SD =1.525,

range =0–14). We also include controls that reﬂect

the history and interaction of the borrower and loan

ofﬁcer to reﬂect their relationship and interpersonal

learning, and to account for other relationship-speciﬁc

effects. Measures include whether this is the client’s

ﬁrst experience with a loan, which is important given

that for many clients microﬁnance is their ﬁrst expo-

sure to ﬁnancial products (mean =0.232, SD =0.422);

whether a loan was restructured, which might signal

that a client has experienced exogenous difﬁculties

(mean =0.006, SD =0.074); the number of loan cycles

the client has had with the MFI, as a proxy for experi-

ential learning (mean =5.701, SD =5.154, range 1–40);

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Table 2 Variables and Descriptive Statistics

Variable Description Mean SD Min Max

BusinessLoan Dummy variable. Takes the value of 1 for incorporated ﬁrms and 0 for individual borrowers

with unregistered businesses.

00000 00016 0 1

ãLoanOfﬁcer Dummy variable. Takes the value of 1 if the loan ofﬁcer was rotated during the loan cycle. 00253 00435 0 1

ClientTenure Number of loan cycles the client has had with the ﬁrm. 50701 50154 1 40

Female Dummy variable. Takes the value of 1 for female clients. 00624 00484 0 1

FirstLoan Dummy variable. Takes the value of 1 if this is the client’s ﬁrst loan cycle with the ﬁrm. 00232 00422 0 1

GroupLoan Dummy variable. Takes the value of 1 for group loans. 00150 00357 0 1

InterestRateaYearly interest rate charged. 800122 60475 58 96

LoanAmount aSize of the original loan, in thousand pesos. 80925 80561 0.50 50

PastDelinquency Total number of previous loan cycles where the client has missed a payment. 10083 10525 0 14

PaymentDueaSize of scheduled payments, in thousand pesos. 10114 10924 0 20

PaymentFrequency Days between scheduled loan payments. 150545 140423 7 86

PreviousRegime Length of relationship, in number of loan cycles, between the client and the previous loan

ofﬁcer—conditional on a changed ofﬁcer.

20370 20196 1 27

RestructuredLoan Dummy variable. Takes the value of 1 if the loan has been restructured. 00006 00074 0 1

ThreeMissedPayments Dummy variable. Takes the value of 1 if a third payment is missed. 00115 00319 0 1

TwoMissedPayments Dummy variable. Takes the value of 1 if a second payment is missed. 00168 00374 0 1

OneMissedPayment Dummy variable. Takes the value of 1 if there has been a missed payment in the loan cycle. 00333 00471 0 1

IndividuallyConsistenttDummy variable. Takes the value of 1 if loan ofﬁcer has a spirit or letter of the law

relational style.

00665 00472 0 1

AlongConsistentt1 t +1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had a spirit or letter of the

law relational style, the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer

at time t+1 had a spirit of the law style if the prior loan ofﬁcer had that style and a

letter of the law style if the prior loan ofﬁcer had that style.

00021 00143 0 1

IndividuallyConsistentt→

IndividuallyConsistentt+1

Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had a spirit or letter

relational style, the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at

time t+1 had either a spirit or letter style.

00019 00136 0 1

IndividuallyInconsistentt→

IndividuallyConsistentt+1

Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had a mixed relational

enforcement style, the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer

at time t+1 had either a spirit or letter of the law style.

00062 00240 0 1

IndividuallyConsistentt→

IndividuallyInconsistentt+1

Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had a spirit or letter of the

law relational style, the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer

at time t+1 had a mixed style.

00059 00235 0 1

SLtDummy variable. Takes the value of 1 if loan ofﬁcer has a spirit of the law relational style. 00352 00478 0 1

LLtDummy variable. Takes the value of 1 if loan ofﬁcer has a letter of the law relational style. 00313 00464 0 1

MtDummy variable. Takes the value of 1 if loan ofﬁcer has a mixed relational style. 00336 00472 0 1

SLt→LLt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had an SL relational style,

the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1 had an

LL style.

00028 00166 0 1

LLt→SLt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had an LL relational style,

the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1 had an

SL style.

00019 00135 0 1

Mt→SLt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had a mixed relational

style, the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1

had an SL style.

00033 0018 0 1

Mt→LLt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had a mixed relational

style, the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1

had an LL style.

00024 00153 0 1

LLt→LLt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had an LL relational style,

the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1 also

had an LL style.

0002 0014 0 1

SLt→SLt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had an SL relational style,

the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1 also

had an SL style.

00032 00177 0 1

Mt→Mt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had a mixed relational

style, the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1

had a mixed style.

00041 00198 0 1

SLt→Mt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had an SL relational style,

the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1 had a

mixed style.

00032 00175 0 1

LLt→Mt+1Dummy variable. Takes the value of 1 if at time tthe loan ofﬁcer had an LL relational style,

the ofﬁcer was reassigned, and the subsequent/different loan ofﬁcer at time t+1 had a

mixed style.

00024 00154 0 1

aThe log of these variables is used in the analyses.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

as well as the number of loan cycles a client had

with a particular loan ofﬁcer, to control for the per-

sonal bond that might have developed between them

(mean =2.370, SD =2.196, range 1–27). Branch and

year ﬁxed effects are included to absorb institutional

and temporal variation. Loan ofﬁcer ﬁxed effects are

also included to ensure that the sequential consistency

effects we observe are not attributable to some unob-

served, time-invariant loan ofﬁcer or borrower char-

acteristics. (Models without loan ofﬁcer ﬁxed effects

yield similar conclusions.)

2.3. Analytical Strategy and Identiﬁcation

We model the probability of late payment with the

following summary model:

Pr8Y =19=G0+

p=1

14IndividuallyConsistentt5

+24ãLoanOfﬁcert1 t+15

+34AlongConsistentt1 t+15

+44IndividuallyInconsistentt

→IndividuallyInconsistentt+15

+54IndividuallyConsistentt

→IndividuallyConsistentt+15

+64X5+äl+i+t+1

where Yis a binary variable denoting that, contrary

to contractual provisions, the borrower missed a

payment; ãLoanOfﬁcert1 t+1denotes a change in loan

ofﬁcer between time tand t+1; IndividuallyConsistentt

is a dummy variable equal to 1 if the loan ofﬁcer orig-

inating the loan had either consistent relational styles

(SLtor LLt5;AlongConsistentt1 t+1is a dummy variable

equal to 1 if the former and subsequent loan ofﬁcers

have the same relational style, e.g., LLt→LLt+1or

SLt→SLt+1(and is therefore contingent on a loan

ofﬁcer change as are the measures that follow);

IndividuallyInconsistentt→IndividuallyInconsistentt+1

is a dummy variable equal to 1 if both the orig-

inating and subsequent loan ofﬁcers employ

mixed relational styles; IndividuallyConsistentt→

IndividuallyConsistentt+1is a dummy variable equal

to 1 if both the originating and subsequent loan ofﬁ-

cers have individually consistent relational styles but

employ different ones (e.g., SLt→LLt+1,LLt→SLt+15

(the omitted category thus represents changes from

individually consistent to individually inconsistent;

e.g., SLt→Mt+1,LLt→Mt+15;Xis a vector of

controls; äldenotes loan ofﬁcer ﬁxed effects; i

denotes branch ﬁxed effects, tdenotes year ﬁxed

effects; and denotes the error term. We also specify

style-speciﬁc models that separate the LL and SL

effects in a similar fashion:9

Pr8Y =19=G0+

p=1

14LLt5+24SLt5

+34ãLoanOfﬁcert1 t+15

+44SLt→LLt+15+54LLt→SLt+15

+64Mt→SLt+15+74Mt→LLt+15

+84SLt→SLt+15+94LLt→LLt+15

+104X5+äl+i+t+0

To ensure that clustering is adequately addressed, we

also estimated the model using a hierarchical nonlin-

ear framework and various standard error clustering

techniques. Results are consistent across models.10

3. Presentation of Findings

We begin by providing bivariate statistics in Figure 1

to build intuition and establish a baseline for the

multivariate analyses that follow. As is evident when

there is a change in loan ofﬁcer, 38.8% of loans expe-

rience a missed payment as opposed to 31.3% when

there is no change, which reﬂects an increase of 24%

(contrast =0.074: t-test =13001, p < 00000); for two or

more missed payments, the difference is even greater

at 47.2% (contrast =0.071: t-test =13092, p < 00000),

and for three or more missed payments, the difference

is 61.3% (contrast =00061: t-test =13042, p < 00000).

Note that all these tests and those that follow are con-

ducted with robust standard errors clustered at the

loan ofﬁcer level. It is thus evident that, as predicted

by relational embeddedness and the literature on

change in organizations, borrowers experience signiﬁ-

cant disruption when the tie to their current loan ofﬁ-

cer is broken. These quantitative ﬁndings are echoed

in our qualitative work. An SL ofﬁcer explained this

common concern: “Transfers are tricky. You can lose a

lot of clients, because the clients are used to working

with a different loan ofﬁcer. So when you arrive they

can be like, ‘Who the f—k are you?’ ” The question is

whether the disruption arises solely from the loss of

a personal relationship or reveals a deeper pattern.

In Figure 2we introduce loan ofﬁcer relational

styles. The bivariate results reveal signiﬁcant varia-

tion by relational style. For loans administered by

ofﬁcers with a mixed style, the percentage with a

9In loan ofﬁcer ﬁxed effects models, 1and 2cannot be identiﬁed.

Models without ﬁxed effects and thus main relational style effects

yield similar results.

10 Given the size and complexity of our models, convergence took

a considerable amount of time, making it impractical for more than

a robustness check.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Figure 1 A Change in Loan Ofﬁcer Results in a Higher Rate of Missed

Payments



























.OCHANGE #HANGE

Source. Data obtained from a unique data set of ≈4501000 microﬁnance

loans made in Mexico, 2004–2008.

Notes. Black bars denote one missed payment; gray bars denote two or more

missed payments. Differences are statistically signiﬁcant at p < 0001. Models

are unconditional.

missed loan payment is 35%, with 19.8% missing two

or more. By contrast, 30.3% of loans administered by

letter of the law ofﬁcers are not paid on time (13.8%

two or more times). Spirit of the law ofﬁcers exhibit

similar patterns as letter of the law ofﬁcers with ﬁg-

ures of 31.8% and 13.3%, respectively. For one missed

payment the contrasts between mixed and letter of

the law (contrast =0.047: t-test =3078, p < 00000) and

mixed and spirit of the law (contrast =0.033: t-test =

2048, p < 00013) are both statistically signiﬁcant, while

the contrast between letter of the law and spirit of the

law is not statistically signiﬁcant (contrast = −00014:

t-test = −0098, p < 00328). (The same statistical results

hold true for two or more or three or more delin-

quencies; available upon request.) These results sug-

gest that borrowers are less likely to miss a payment

if their loan ofﬁcers have an individually consistent

style, irrespective of whether this style entails strict

interpretation and enforcement of the rules or a more

ﬂexible style. Following this intuition, we present

Figure 2 Missed Payment Rates Vary by Loan Ofﬁcer Relational

Enforcement Styles

32% 30%

35%

13% 14%

20%

Spirit Letter Mixed

One Two or more

Source. Data were obtained from a unique data set of ≈4501000 microﬁ-

nance loans made in Mexico between 2004 and 2008.

Notes. The black line denotes one missed payment; the gray line denotes

two or more missed payments. Differences are statistically signiﬁcant at p <

0001. “Spirit” denotes spirit of the law loan ofﬁcers, “Letter” denotes letter

of the law loan ofﬁcers, and “Mixed” denotes loan ofﬁcers who blend and

vacillate between styles. Models are unconditional.

ﬁndings integrating both consistent styles as well as

comparing each style separately.

To provide intuition concerning inter-loan ofﬁcer

changes, Figure 3(a) illustrates unconditional loan

ofﬁcer change and missed payment rates by sum-

mary categories of relational styles, and Figure 3(b)

breaks out results by speciﬁc transitions within or

between speciﬁc styles. The ﬁrst two columns provide

baselines for one or two missed payments when there

is no change in loan ofﬁcer. In such cases, 31.4% of

loans experience a missed payment, with 15% experi-

encing two or more. The proportions are considerably

Figure 3 Missed Payment Rates Vary by Transitions Between and

Within Loan Ofﬁcer Relational Enforcement Styles

31% 33% 35%

37% 40%

44%

15% 15% 17% 18%

25%

28%

(a)

(b)

No change Inconsistent

Consistent

Along-

Consistent

Inconsistent

Consistent

Inconsistent

33% 34% 35% 35% 37% 38% 40%

43% 46%

16% 15% 16% 18% 16%

20%

25% 27% 28%

M→LL M→SL SL→SL LL→LL LL→SL SL→LL M→MSL→MLL→M

→→→→

Source. Data were obtained from a unique data set of ≈4501000 micro-

ﬁnance loans made in Mexico between 2004 and 2008, by an urban-

focused MFI.

Notes. The black line denotes the sample average rate for one missed pay-

ment if there is a change in loan ofﬁcer; the grey line, two missed payments.

Dark vertical bars represent one missed payment; variegated lines represent

two missed payments. Statistics presented are unconditional. AlongConsis-

tent denotes successive loan ofﬁcers with the same individually consistent

styles (i.e., SL →SL,LL →LL), Consistent →Consistent denotes mov-

ing between loan ofﬁcers with individually consistent but different styles

(i.e., LL →SL,SL →LL), Inconsistent →Inconsistent denotes successive

loan ofﬁcers each with individually inconsistent styles, and Consistent →

Inconsistent denotes moving from a loan ofﬁcer with an individually consis-

tent style (spirit or letter) to one with a mixed style. SL denotes spirit of the

law loan ofﬁcers, LL denotes letter of the law loan ofﬁcers, and Mdenotes

loan ofﬁcers who vacillate between styles. Arrows denote that, contingent

on a change in loan ofﬁcer, the prior loan ofﬁcer had the ﬁrst enforcement

style and the subsequent one the style after the arrow. For example, M→LL

denotes that the loan ofﬁcer at time thad a mixed enforcement style and the

subsequent loan ofﬁcer, at time t+1, had a letter of the law enforcement

style. Models are unconditional.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

higher when there is a change of any kind, as previ-

ously noted. However, there is signiﬁcant variation in

these ﬁgures depending on the pattern of loan ofﬁcer

relational style transitions. For example, when the for-

mer loan ofﬁcer had a mixed style and the subsequent

one an individually consistent style (either spirit or

letter), the missed payment rate is approximately 33%

for one missed payment and 15% for two. These rates

are roughly 14% and 31% less than the overall aver-

age for an ofﬁcer change (indicated by the horizontal

bars in the graphs). Moreover, for two missed pay-

ments, the ﬁgure is the same as that for the situation

in which there is no change. It is thus evident that the

style employed by subsequent loan ofﬁcer matters.

The columns labeled “AlongConsistent” show that

when the prior and randomly assigned subsequent

loan ofﬁcers employ the same relational style, there

is a 9% reduction in the percentage of loans missing

a payment and a 23% reduction in two missed pay-

ments. Reductions are also evident with transitions

across individually consistent styles (spirit to letter, or

vice versa). Increases are evident when there are tran-

sitions within mixed styles (+3% and +14% for one

or two missed payments, respectively), with the worst

outcome arising from a move from a loan ofﬁcer with

an individually (internally) consistent style (spirit or

letter) to one who employs a mixed style (+14% and

+25% for one or two missed payments, respectively).

Figure 3(b) plots changes within and between spe-

ciﬁc relational styles. When there is a change in loan

ofﬁcer, but the old and new loan ofﬁcers both employ

a mixed style, 40% experience one missed payment

and 25.1% two or more. For letter of the law ofﬁ-

cers, the ﬁgures are 35.3% and 18.2%, and the cor-

responding ﬁgures for spirit of the law ofﬁcers are

35% and 16.2%. Although a change in loan ofﬁcer

appears to be detrimental in general,11 it is consid-

erably more detrimental when the loan ofﬁcers have

mixed styles, as they result in 54.9% more multi-

ple delinquencies than changes in spirit of the law

ofﬁcers (contrast =0005: t-test =2022, p < 00026) and

38.2% more than a comparable change in letter of the

law ofﬁcers (contrast =00048: t-test =2046, p < 00014).

A movement from either a spirit of the law (43.2% for

one and 26.5% for two or more missed payments) or

letter of the law (46% for one and 28% for two or more

missed payments) relational style to one with a mixed

style results in a signiﬁcantly greater percentage of

missed loan payments (contrasts are all statistically

signiﬁcant). Moving from a mixed style to either an

LL (32.5% for one and 15.8% for two or more missed

payments) or SL (33.7% for one and 14.7% for two or

11 All contrasts are statistically signiﬁcant when compared with no

change.

more missed payments) relational style decreases this

percentage appreciably.

Unpacking the results further, it appears that mov-

ing from a mixed to a letter of the law loan ofﬁcer

reduces the probability of a missed payment by more

than 41% (contrast = −0013: t-test = −6074, p < 00000);

this compares with a 28% reduction for similar move-

ments from mixed to spirit of the law (contrast =

−00095: t-test = −5039, p < 00000). The results for two

or more missed payments are even starker. But, inter-

estingly, here the largest comparative difference is

between moving from a mixed loan ofﬁcer to one

with a spirit of the law style (85.7%) rather than vice

versa. An additional pattern is worth highlighting:

notice that changes to SL and LL styles perform sim-

ilarly for the ﬁrst missed payment. At the same time,

clients who are assigned to SL ofﬁcers after a rotation

are signiﬁcantly less likely to miss their second and

third payments, practically eliminating the impact of

a broken tie. Thus, regardless of their preceding expe-

rience, clients who are assigned to SL ofﬁcers consis-

tently “recover” from a change more promptly.

Bivariate analyses thus far reveal that borrow-

ers exhibit signiﬁcantly different repayment patterns

based on the consistency of their loan ofﬁcers’ rela-

tional styles. This implies that relational styles are

consequential, as borrowers are sensitive to them

regardless of their formal contractual responsibilities

or the power dynamics underlying their relationships

with loan ofﬁcers. To ensure that these results are not

spurious, we present several multivariate models that

account for the characteristics of the borrower, the

loan itself, and the institutional and economic envi-

ronments within which loans are made below to test

our hypotheses.

Table 3presents logistic regression coefﬁcients pre-

dicting the likelihood of breaching contractual provi-

sions concerning timely loan repayment. The models

control for the natural log of interest rate charged,

frequency of loan payments, natural log of the total

amount of loan, gender of the borrower, whether the

loan was received as part of a group, and whether

the borrower has a history of late payment. The con-

trols are all highly statistically signiﬁcant and in the

anticipated directions, and they are included in all

models.

We build intuition by ﬁrst presenting a summary

model to demonstrate aggregate consistency effects.

Model 1 reveals that loans administered by an ofﬁ-

cer employing an individually consistent style (either

letter or spirit) are less likely to experience a missed

payment than is the case with a mixed style, which is

the omitted category (eb4−001255=0088, p < 00001 (two-

tailed test, as are all that follow)). Model 2 separates

the summary individually consistent effect into its

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Table 3 Logistic Regression Predicting a Missed Payment in a Loan Cycle

1234567

Variable b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE)

IndividuallyConsistentt(H1) −00125

4000075∗∗∗

SLt(H1) −0013 −00079

4000095∗∗∗ 4000095∗∗∗

LLt(H1) −00118 −00091

4000105∗∗∗ 4000105∗∗∗

Mt(H1) 00125

4000075∗∗∗

ãLoanOfﬁcert1 t +100427 00419 00379 00187

4000095∗∗∗ 4000195∗∗∗ 4000155∗∗∗ 4000165∗∗∗

AlongConsistentt1 t +1(H2) −00206

4000305∗∗∗

IndividuallyConsistentt→ −00183

IndividuallyConsistentt+1(H2) 4000315∗∗∗

IndividuallyInconsistentt→ −00066 00165

IndividuallyInconsistentt+1(H2) 4000195∗∗∗ 4000195∗∗∗

IndividuallyInconsistentt→ −00251

IndividuallyConsistentt+1(H3) 4000235∗∗∗

SLt→SLt+1(H2) −00178

4000345∗∗∗

LLt→LLt+1(H2) −0015

4000435∗∗∗

SLt→LLt+1(H2) −00145

4000365∗∗∗

LLt→SLt+1(H2) −00141

4000435∗∗∗

Mt→SLt+1(H3) −00214

4000235∗∗∗

Mt→LLt+1(H3) −00206

4000285∗∗∗

SLt→Mt+1(H3) 0019

4000245∗∗∗

LLt→Mt+1(H3) 00284

4000275∗∗∗

Loan ofﬁcer ﬁxed effects NO NO NO NO YES YES YES

Branch ﬁxed effects YES YES YES YES YES YES YES

Year ﬁxed effects YES YES YES YES YES YES YES

Model ﬁt/diagnostics

N438,252 438,252 438,252 438,252 438,346 438,346 438,346

248,031∗∗∗ 48,030∗∗∗ 48,031∗∗∗ 49,998∗∗∗ 46,402∗∗∗ 46,385∗∗∗ 46,395∗∗∗

Notes. Data were obtained from a proprietary, loan-level database of microﬁnance loans from one urban-market-focused MFI in Mexico between 2004 and

2008. The dependent variable is dichotomous and takes the value of 1 if, within a loan cycle, a client has missed a payment. In addition, all models include

the following controls, which were omitted from the table for presentation purposes and because the coefﬁcients are remarkably stable across speciﬁcations

(coefﬁcients signiﬁcant at the 0.01 level in parentheses): PaymentFrequency 4−000215, ln(LoanAmount)(0.817), ln(PaymentDue)4−007775, ln(InterestRate)

(2.09), Female 4−000255,GroupLoan (0.118), ClientTenure 4−001375,BusinessLoan (0.462), PastDelinquency (0.417), RestructuredLoan (1.214), FirstLoan

(0.38), and PreviousRegime (0.025). The omitted category for models 1 and 2 is a Mixed style. For model 3, the omitted category is the individually consistent

category; for model 5, the omitted category is individually consistent to individually inconsistent; in model 6, it is spirit or letter to mixed; and in model 7, it

reﬂects any style-speciﬁc movement from mixed to something else, or between LL and SL.

∗∗∗p < 00001; ∗∗ p < 0001; ∗p < 0005 (two-tailed tests).

constituent spirit (eb4−00135=00878, p < 00001) and let-

ter (eb4−001185=00889, p < 00001) components, reveal-

ing that both tend to outperform the omitted baseline

mixed style. Model 3 includes a measure for a mixed

(intra-loan ofﬁcer) style. The coefﬁcient is positive,

suggesting that relative to either an LL or SL loan

ofﬁcer (the omitted category), those loans adminis-

tered by an individually inconsistent loan ofﬁcer are

more likely to experience a missed payment. These

tests individually and collectively provide support

for the ﬁrst hypothesis that loans administered by

loan ofﬁcers with individually (internally) consistent

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

relational styles are less likely to miss a payment con-

trary to contractual terms.

Model 4 demonstrates what happens when a loan

ofﬁcer changes: there is a signiﬁcant increase of more

than 50% in the odds of delinquency (eb4004275=1053,

p < 00001). The effect is even stronger for two or

more missed payments (model 3, Table A2: eb4008115=

2025, p < 00001) or three or more (model 3, Table A3:

eb400100255=2079, p < 00001). That these disruptive

effects increase with the length of the relationship a

client had with her previous loan ofﬁcer is clear evi-

dence of relational considerations between clients and

their loan ofﬁcers (eb4000255=1003 for every additional

loan cycle, up to a maximum of 27).

Model 5—including loan ofﬁcer ﬁxed effects (LO

FEs)—begins to provide a test of hypothesis two

concerning inter-loan ofﬁcer consistency in relational

styles. Compared with a movement from an individ-

ually consistent to a mixed style, transitioning from

a loan ofﬁcer employing an individually consistent

style (e.g., spirit, letter) to another ofﬁcer employing a

different style that is nonetheless individually consis-

tent mitigates the impact of severing a tie, reducing

the negative shock from 52% to 26.6% (eb400419−001835,

p < 00001), a result ampliﬁed in models of two or three

missed payments. Transitions along consistent styles

reduce the negative shock of a loan ofﬁcer change

from 52% to 23.7% (eb400419−002065,p < 00001), or a 54%

reduction as predicted in Hypothesis 2. Indeed, even

movements along mixed styles moderate the impact

of change (eb400419−000665=10423 (versus 1.52), p < 00001),

a result that holds across models. Similar results are

evident for two (model 5, Table A2) or three (model 5,

Table A3) missed payments.12

Model 6 provides estimates for style-speciﬁc tran-

sitions (e.g., LLt→LLt+15. The overall results are

consistent with the summary measures: relational

style-consistent moves—particularly those made

between spirit of the law ofﬁcers—tend to mitigate

the impact of change (eb400379−001785,p < 00001), as do

movements across different but internally consistent

forms (e.g., letter to spirit), albeit at somewhat

reduced rates (eb400379−001415,p < 00001).13

Hypothesis 3 predicted that borrowers are less

likely to miss a payment when there is a change in

their loan ofﬁcer and the subsequent loan ofﬁcer has

a different style that is also individually consistent.

12 All contrasts are statistically signiﬁcantly different in models

without LO FEs, and all are signiﬁcant in those with LO FEs save

for the differences between mixed to individually consistent and

along consistent styles versus across consistent. All correspond-

ing contrasts are statistically signiﬁcantly different for two or three

missed payments.

13 Differences are statistically signiﬁcant in models without loan

ofﬁcer ﬁxed effects. All corresponding contrasts are statistically sig-

niﬁcantly different for two or three missed payments.

Style-speciﬁc results in model 7 provide strong sup-

port for this argument. Compared with transitions

that are along consistent styles (e.g., LLt→LLt+15

or across different but individually consistent forms

(LLt→SLt+15, transitions from a consistent style to a

mixed style increase the odds of one, two, or more

missed payments appreciably. This is especially true

when moving from a letter of the law to mixed style,

which results in a 60% increase in the odds of missing

a payment (4eb400187+002845−15∗100, p < 00001). These

results also hold across model speciﬁcations (see mod-

els 6 and 6(b) in Tables A2 and A3).

3.1. Robustness Checks

To ensure that the ﬁndings are not driven by client

or loan ofﬁcer subpopulations, we performed a host

of robustness checks (some of these are presented in

Table 4, and all others are available upon request).

All robustness checks include loan ofﬁcer, branch, and

year ﬁxed effects. The ﬁrst two models, 8 and 9, are

restricted to female and male borrowers, respectively,

to determine whether missed payments vary by bor-

rower gender. There is no evidence of this. More-

over, coefﬁcient estimates tend to be similar for both.14

Models 10 and 11 are restricted to group and indi-

vidual borrowers, respectively. Results indicate that

our hypothesized mechanisms operate for both, con-

sistent with work by Giné and Karlan (2012). Mod-

els 12 and 13 are restricted to clients with less than

and more than average tenure with the ﬁrm, respec-

tively, to assess differences in possible learning effects.

Models 14 and 15 are contingent on loan sizes below

and above the median, respectively, to proxy for client

wealth, since loan sizes are correlated with it. Mod-

els 16 and 17 distinguish the gender of the loan ofﬁcer

to determine whether style effects vary as a conse-

quence of whether a male or female employs each.

Finally, models 18–20 are conditional on a loan ofﬁ-

cer change (rather than based on interaction terms).

Furthermore, to ensure that the results are not driven

by unobservable loan ofﬁcer skill, we restricted mod-

els to loan ofﬁcers who, in the organization’s perfor-

mance assessment, performed above the median, in

the top quartile, or in the two mid-quartiles, between

the 25th and 75th percentiles (results omitted to con-

serve space but are available upon request). The

mechanisms we outline hold.

Contextualizing all the previous ﬁndings, it is

important to note that when a new loan ofﬁcer is

assigned to administer a loan, the borrower does not

choose the loan ofﬁcer’s identity or relational style.

14 Statistically comparing across models is a nontrivial challenge.

Because our hypotheses are agnostic to sample differences, and the

point of these checks is to establish robustness rather than theorize

differences, we refrain from such comparisons here.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Table 4 Robustness Checks: Logistic Regression Predicting a Missed Payment in a Loan Cycle

8 9 10 11 12 13 14 15 16 17 18 19 20

Variable b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE) b(SE)

ãLoanOfﬁcert1 t +100523 00516 00673 0048 00517 00435 00538 00481 00504 0055

4000225∗∗∗ 4000295∗∗∗ 4000445∗∗∗ 4000195∗∗∗ 4000315∗∗∗ 4000285∗∗∗ 4000265∗∗∗ 4000255∗∗∗ 4000295∗∗∗ 4000275∗∗∗

AlongConsistentt1 t +1−00275 −00217 −00266 −00242 −0026 −00331 −00314 −00213 −00187 −00344 −00663

4000365∗∗∗ 4000465∗∗∗ 4000755∗∗∗ 4000315∗∗∗ 4000375∗∗∗ 4000465∗∗∗ 4000425∗∗∗ 4000395∗∗∗ 4000425∗∗∗ 4000425∗∗∗ 4000375∗∗∗

IndividuallyInconsistentt→ −00332 −00338 −00363 −00317 −00289 −00433 −00359 −0031 −00374 −00361 −00714

IndividuallyConsistentt+14000275∗∗∗ 4000345∗∗∗ 4000525∗∗∗ 4000235∗∗∗ 4000285∗∗∗ 4000335∗∗∗ 4000305∗∗∗ 4000305∗∗∗ 4000345∗∗∗ 4000305∗∗∗ 4000315∗∗∗

IndividuallyConsistentt→ −00211 −00228 −0023 −00212 −00247 −00233 −00197 −00243 −00308 −00219 −00659

IndividuallyConsistentt+14000375∗∗∗ 4000475∗∗∗ 4000725∗∗∗ 4000325∗∗∗ 4000385∗∗∗ 4000475∗∗∗ 4000425∗∗∗ 4000415∗∗∗ 4000455∗∗∗ 4000415∗∗∗ 4000395∗∗∗

Mt→Mt+1−00048 −00084 −00082 −00054 −00055 −0007 −00064 −00053 −0002 −00044 −00106 00589

4000225∗∗ 4000285∗∗∗ 4000415∗∗ 4000195∗∗∗ 4000245∗∗ 4000275∗∗∗ 4000255∗∗∗ 4000245∗∗ −00029 −00027 4000215∗∗∗ 4000285∗∗∗

SLt→SLt+1−00606

4000465∗∗∗

SLt→LLt+1−00611

4000515∗∗∗

LLt→LLt+1−00616

4000555∗∗∗

LLt→SLt+1−00595

4000545∗∗∗

Mt→SLt+1−00649

4000375∗∗∗

Mt→LLt+1−00662

4000425∗∗∗

SLt→Mt+100645

4000325∗∗∗

LLt→Mt+100763

4000355∗∗∗

Loan ofﬁcer ﬁxed effects YES YES YES YES YES YES YES YES YES YES YES YES YES

Branch ﬁxed effects YES YES YES YES YES YES YES YES YES YES YES YES YES

Year ﬁxed effects YES YES YES YES YES YES YES YES YES YES YES YES YES

Sample/Restrictions Female Male Groups Individual Tenure <Tenure >Size of Size of LO LO Conditional on loan

borrowers borrowers borrowers Med. Med. payment <Med. payment >Med. female male ofﬁcer change

Model ﬁt

N273,268 164,877 63,624 374,548 232,751 205,484 213,662 224,590 174,290 218,454 112,188 112,188 112,188

231,123∗∗∗ 18,894∗∗∗ 5,763∗∗∗ 43,985∗∗∗ 19,087∗∗∗ 26,748∗∗∗ 23,284∗∗∗ 27,549∗∗∗ 21,187∗∗∗ 24,897∗∗∗ 12,058∗∗∗ 12,037∗∗∗ 12,063∗∗∗

Notes. Data were obtained from a proprietary, loan-level database of microﬁnance loans from one urban-market-focused MFI in Mexico between 2004 and 2008. The dependent variable is dichotomous and takes

the value of 1 if, within a loan cycle, a client has missed a payment. Where indicated, models also include loan ofﬁcer (LO) ﬁxed effects. In addition, all models include the controls outlined in the previous table.

∗∗∗p < 00001; ∗∗ p < 0001; ∗p < 0005 (two-tailed tests).

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

This exogenous change in loan ofﬁcer style contin-

gent on change enables us to identify to what extent

moving from an inconsistent relational style to a con-

sistent one impacts breach of contract in the form of

a missed payment. The results suggest that moving

from a mixed to a spirit of the law relational enforce-

ment style reduces the odds of delinquency apprecia-

bly. The same is true for a movement from a mixed to

a letter of the law relational style. The converse is also

true: moving from an individually consistent style

(either LL or SL) to an inconsistent style increases the

odds of a delinquency (see model 2). Moreover, the

results are ampliﬁed when we consider two or more

or three or more delinquencies.15

Finally, it becomes clear that clients who are trans-

ferred to an LL or SL ofﬁcer have similar likelihoods

of missing a ﬁrst payment, but across the board, we

ﬁnd that clients who are transferred to an SL ofﬁcer

have a signiﬁcantly lower likelihood of missing their

second and third payments. The pattern may seem

puzzling, but together with our qualitative data, it

provides evidence of the mechanisms that we theo-

rized affect loan outcomes after a broken tie. In dis-

cussing why consistency should matter in exchange

relationships, we argued that loan ofﬁcers “educate”

clients not only on the technical aspects of a loan but

also on the relational expectations of a lending rela-

tionship. When the patterns that clients expect in their

loan interactions change, clients experience a disrup-

tion in the relational contract with the MFI. But the

replacement loan ofﬁcer can mitigate this by taking

the time to “reeducate” existing clients and establish

new relational expectations:

Before I move to my new branch I go out with the

loan ofﬁcer and ask him to take me to his clients and

introduce me, and then I spend time explaining to the

clients how I like to work and letting them ask me all

their questions. Then when I am transferred I go out

again, and I say, “Remember me? I am your new loan

ofﬁcer,” and I tell them again how my style is different

and why I think it is better. (SL ofﬁcer)

From this perspective, it is less surprising that SL

ofﬁcers eliminate the negative impact of a broken tie

by the second and third missed payment of a loan.

In loan ofﬁcer interviews and observations, SL ofﬁ-

cers not only spoke more often and in more depth

about the importance of educating clients, but we also

observed them spend much more time in client edu-

cation activities, especially following a branch rota-

tion. In fact, it is an inherent characteristic of the SL

style that loan ofﬁcers spend more time exchanging

15 We estimated the model using a two-level (Bernoulli) hierarchical

model. The results (available upon request) were substantively and

statistically similar.

soft information with their clients. This includes infor-

mation on why clients miss payments if their loan

ofﬁcers are transferred and on the expectations they

created during their previous lending relationships

that no longer hold. This pattern provides additional

insights into the implications of our ﬁndings for orga-

nizations, which we discuss in the next section.

4. Discussion and Conclusion

Formal contracts constitute an important mechanism

to facilitate repeated exchange, which is essential in

social and economic life. Given inherent limitations in

contracts, it has long been shown that organizations

can derive signiﬁcant value through relationships,

theorized as relational contracts or as the embed-

ded ties that enable them. However, an organization

cannot establish these relationships—only individuals

authorized to do so on its behalf can. It is thus unclear,

considering the inevitability of turnover and change,

when and how organizations can retain the value cre-

ated through relational mechanisms (Sorenson and

Rogan 2014). Put differently, we need more clarity on

how relational ties, which are established and main-

tained by individuals, can become an organizational

capability.

We propose theory that speciﬁes how both within-

and between-employee consistency in relational styles,

net of relational embeddedness, allow an organization

to largely retain the value of its employees’ relational

ties. We deﬁne relational styles as discernible, reoc-

curring patterns of interaction within and between

social actors. Informed by rich qualitative and quan-

titative evidence, we demonstrate that loan ofﬁcers

who employ consistent relational styles improve loan

outcomes. We also show that when a loan ofﬁcer

leaves the ﬁrm, replacing him with another ofﬁcer

who employs a consistent relational style mitigates

the negative effects of the broken tie.

We use this evidence to argue that organizations

can derive signiﬁcant and sustainable advantage by

developing capabilities to establish consistent pat-

terns of interaction with their clients, which in turn

allow them to retain the value of the relational ties

otherwise anchored in their employees. By implica-

tion, speciﬁc people or positions become less impor-

tant, and consistent relational styles can constitute an

important link between formal organizational prac-

tices and actual client experiences. This ﬁnding may

be particularly valuable in settings or situations where

organizations assign client relationships to a sin-

gle employee, rendering the organization particularly

susceptible to employee turnover.

This implies that organizations should establish

routines that, in alignment with their identity, select

and socialize employees to develop consistent rela-

tional styles with clients. They should also train

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

employees to “educate” clients on the relational

expectations of their interactions with the organiza-

tion. To the extent that an organization can maintain

consistent relational styles with clients across indi-

vidual employees, it will be able to establish clear

and predictable relational expectations, thus increas-

ing its ability to appropriate the value of relational ties

despite employee turnover and exogenous change.

The ﬁndings also have sociological implications.

Prior theory has considered the impact individuals (or

dyadic ties) and positions (e.g., roles) have on a wide

range of outcomes of interest. For example, it is well

documented that organizations gain value through

the social capital generated by interpersonal ties. At

the same time, relational value travels through differ-

ent mechanisms, with implications for the ﬁrm’s (ver-

sus the employee’s) ability to appropriate it (Sorenson

and Rogan 2014). In this research we provide addi-

tional speciﬁcity on the sources of relational value and

an organization’s ability to appropriate it. We demon-

strate that the “logic of affection” plays a critical role

but is retained by individual employees.

Yet we also uncover a different layer of social struc-

ture that plays a larger role and is easier for the

organization to appropriate: typologies of relational

styles within and between social actors that establish

clear and stable relational expectations, particularly

in the face of change. Whereas relational embedded-

ness and the network literature more generally focus

on what social ties look like and their implications,

relational styles concern how interpersonal ties are

enacted and their implications for repeated exchange.

Whereas individuals own relational ties, relational

styles can become an organizational routine and, as a

result, transfer ownership of the relational value to the

organization.

Relational styles may also inform other theoreti-

cal concerns. Norm enforcement, for example, is usu-

ally understood to improve with cohesion (structural

embeddedness). We show that compliance to norms

of behavior may also be determined by a shared

understanding of expectations and the social scripts

that underpin them. A change in script (through a

change in style) is problematic because it removes

shared expectations, weakening norm compliance.

This paper suggests the possibility that people who

are similar on several dimensions may also follow

similar relational scripts, making exchange smoother

and more efﬁcient.

As this is an initial statement concerning rela-

tional styles, many questions follow. First, the data

used here are rich and extensive. However, they are

derived from one industry, in one country, during

a speciﬁc period of time. The extent to which these

ﬁndings apply in other settings should be investi-

gated. Second, additional theorization and investiga-

tion into the antecedents, forms, and consequences of

consistency in relational styles is necessary. Indeed,

we believe there is signiﬁcant opportunity to broaden

and deepen our collective understanding of interac-

tion patterns and their consequences in a range of set-

tings by considering relational styles, which we view

as a theoretical concept that integrates and extends

insights from rich literatures concerning categories

and relational embeddedness. This includes the basic

question of how relational styles develop in the ﬁrst

place. Our intuition on the matter, based on ﬁeld

observations, is that the origin of relational styles rests

largely in social structures including one’s family,

early socialization, initial (i.e., imprinting) and subse-

quent (in timing, quality, and duration) employment

experiences and genealogy, as well as more general

cultural forces. In this case, anecdotal evidence sug-

gests that loan ofﬁcers who were recruited because

they were particularly successful clients or commu-

nity leaders and were ﬁrst trained by an SL loan

ofﬁcer tend to become SL loan ofﬁcers themselves.

By contrast, loan ofﬁcers who were recruited out of

college (typically from an economics or ﬁnance major)

and were ﬁrst trained by an LL loan ofﬁcer tend to

develop LL relational styles.16 The relative magnitude

and interaction of each of these forces is ultimately an

empirical question.

One can imagine other social settings where similar

dynamics are at work. Labor and dating markets are

clear examples, as are alliance relationships between

ﬁrms. In each of these cases, learning about how to

interact and transact with another party occurs at one

point in time. But this learning is not necessarily part-

ner or corporate actor speciﬁc. There are types of rela-

tional styles that reoccur, and having experience with

a particular type helps shape future understanding

and action when facing individuals who exhibit simi-

lar relational styles. Moreover, because of path depen-

dencies and self-fulﬁlling dynamics, these relational

style types can remain stable across interactions and

settings (e.g., Schelling 1960). One can imagine, how-

ever, a wide range of types of styles that are concep-

tually nested within a set of principles that serve as

building blocks for variations. We view letter of the

law and spirit of the law types as general categories

of actors’ understandings of the nature of organiza-

tional rules and their enactment, which can be applied

in a host of settings including, for example, educa-

tion, law enforcement, as well as in childrearing or

the sociology of marriage and family. To lower vio-

lence levels in inner-city communities, for example,

police forces must establish trust-based, productive

relationships with the community while also retain-

ing authority and clear territorial control. The “broken

16 These two represent the most common recruitment strategies.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

windows” approach to violence reduction, for exam-

ple, is based on the premise that the police will react

with absolute consistency across all types of criminal

acts, regardless of their magnitude.

Community policing models, by contrast, are based

on the premise that no police force can be effec-

tive without the community’s trust, which is built

by police ofﬁcers who consistently show personal

involvement and concern for community members.

Consistent with this point, Google’s internal research

ﬁnds that whereas managers’ personalities and styles

can vary widely, consistency and predictability with

subordinates emerged as the best predictor of man-

agerial performance (Bryant 2011).

As noted above, many organizations spend consid-

erable time training employees on how to interact with

customers with a speciﬁc focus on providing a con-

sistent customer experience across employees. When

they join “The Happiest Place on Earth,” Disneyland

employees are heavily socialized. But so are visitors

(Van Maanen 1991). These taught and learned rela-

tional styles can be a key element of the organization’s

value proposition. From a customer’s perspective, it

clariﬁes expectations for interactions with the orga-

nization. Expectations are similarly set for employ-

ees concerning appropriate ways of interacting with

clients. Relational styles thus buffer the organiza-

tion against changes in personnel that could dam-

age customer–client relationships because customers

are not tied to organizational employees but rather

to how employees of the organization generally inter-

act with them. Creating organizational capabilities to

establish consistent relational styles may be especially

important for organizations that, similar to MFIs,

have dual or hybrid identities (Battilana and Dorado

2010). In the case of MFIs, key constituencies may

have different expectations for whether the organi-

zation should be primarily committed to social out-

comes or to ﬁnancial returns. A lack of consistency in

how employees relate to different constituencies may

create confusion that can even lead to social unrest.

General principles of interaction can vary across

cultures. A question is thus to what extent relational

styles vary in form and effect across social space. In

our setting, for example, we ﬁnd that clients are reac-

tive to a change in relational style even though they

are relatively dependent on their loans and are in

a situation of clear power disadvantage. We would

expect sensitivity to relational styles to increase with

the relative power of the counterpart. In similar fash-

ion, we see an effect of relational styles even though

loan contracts in microﬁnance are of the simplest

type. We would expect the relevance of relational

styles to increase with the technical complexity of

the information transferred in an exchange as the

necessity for, and volume of, information required

for interpretation increases. The extent to which stan-

dardization is important in one context can be one

basis of variation. In some industries or businesses,

there may be an expectation for individually con-

sistent relational styles but not necessarily across-

actor consistency. A business context that prides itself

on individuality or quirkiness may favor such an

approach. Finally, although we were able to exploit

random assignment of subsequent loan ofﬁcers to test

our predictions concerning inter-loan ofﬁcer inconsis-

tency, the data used here cannot support such strong

inferences with respect to intra-loan ofﬁcer consis-

tency, which would require a different research design

and data structure.

Supplemental Material

Supplemental material to this paper is available at http://dx

.doi.org/10.1287/mnsc.2015.2167.

Acknowledgments

Authorship is alphabetical; both authors contributed

equally in the development of this paper. The authors thank

Gino Cattani, Joe Porac, Catherine Turco, Ezra Zuckerman,

and seminar participants at Chicago, Cornell, Duke, Har-

vard Business School, the Massachusetts Institute of Tech-

nology, New York University, and Yale for their feedback

on earlier drafts. The authors also thank three anonymous

reviewers and the Management Science associate and depart-

ment editors. All errors and omissions are the authors’

alone. Please address correspondence to the second author.

References

Anthony D (2005) Cooperation in microcredit borrowing groups:

Identity, sanctions, and reciprocity in the production of collec-

tive goods. Amer. Sociol. Rev. 70(3):496–515.

Ariely D (2008) Predictably Irrational: The Hidden Forces That Shape

Our Decisions (HarperCollins, New York).

Armendáriz de Aghion B (1999) On the design of a credit agree-

ment with peer monitoring. J. Development Econom. 60(1):

79–104.

Armendáriz de Aghion B, Morduch J (2000) Microﬁnance beyond

group lending. Econom. Trans. 8(2):401–420.

Armendáriz de Aghion B, Morduch J (2004) Microﬁnance: Where

do we stand? Goodhard CAE, ed. Financial Development and

Economic Growth: Explaining the Links (Palgrave Macmillan, Bas-

ingstoke, UK), 135–148.

Arrow K (1972) Models of job discrimination. Pascal A, ed. Racial

Discrimination in Economic Life (D.C. Heath, Lexington, MA),

83–102.

Baker G, Gibbons R, Murphy K (1994) Subjective performance

measures in optimal incentive contracts. Quart. J. Econom.

109(4):1125–1156.

Baker WE, Faulkner RR, Fisher GA (1998) Hazards of the market:

The continuity and dissolution of interorganizational market

relationships. Amer. Sociol. Rev. 63(2):147–177.

Baker G, Gibbons R, Murphy K (2002) Relational contracts and the

theory of the ﬁrm. Quart. J. Econom. 117(1):39–84.

Baldwin M (1992) Relational schemas and the processing of social

information. Psych. Bull. 112(3):461–484.

Baron J, Hannan M, Burton M (2001) Labor pains: Organizational

change and employee turnover in young, high-tech ﬁrms.

Amer. J. Sociol. 106(4):960–1012.

Battilana J, Dorado S (2010) Building sustainable hybrid organiza-

tions: The case of commercial microﬁnance organizations. Acad.

Management J. 53(6):1419–1440.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Beatty S, Morris M, Coleman JE, Ellis K, Lee J (1996) Customer-sales

associate retail relationships. J. Retailing 72(3):223–247.

Becker G (1959) The Economics of Discrimination (University of

Chicago Press, Chicago).

Berger P, Luckmann T (1967) The Social Construction of Reality, A

Treatise in the Sociology of Knowledge (Doubleday, New York).

Berger A, Udell G (1995) Relationship lending and lines of credit

in small ﬁrm ﬁnance. J. Bus. 67(3):351–381.

Berger A, Klapper L, Udell G (2001) The ability of banks to lend

to informationally opaque small businesses. J. Banking Finance

25:2127–2167.

Besley T, Coate S (1995) Group lending, repayment incentives and

social collateral. J. Development Econom. 46(1):1–18.

Bester H (1985) Screening vs. rationing in credit markets with

imperfect information. Amer. Econom. Rev. 75(4):850–855.

Bielby W, Baron J (1986) Men and women at work: Sex segregation,

and statistical discrimination. Amer. J. Sociol. 91(4):759–799.

Biong H, Ulvnes AM (2011) If the supplier’s human capital walks

away, where would the customer go? J. Bus.-to-Bus. Marketing

18(3):223–252.

Broschak J (2004) Managers’ mobility and market interface: The

effects of managers’ career mobility on the dissolution of mar-

ket ties. Admin. Sci. Quart. 49(4):608–640.

Broschak J, Block E (2014) With or without you: When does man-

agerial exit matter for the dissolution of dyadic market ties?-

Acad. Management J. 57(3):743–765.

Bryant A (2011) Google’s quest to build a better boss. New

York Times (March 12), http://www.nytimes.com/2011/03/13/

business/13hire.html.

Burt R (2000) Decay functions. Soc. Networks 22:1–28.

Burt R (2001) Attachment, decay and social network. J. Organ.

Behav. 22(6):619–643.

Burton M, Beckman C (2007) Leaving a legacy: Position imprints

and successor turnover in young ﬁrms. Amer. Sociol. Rev.

72(2):239–266.

Canales R (2011) Rule bending, sociological citizenship, and orga-

nizational contestation in microﬁnance. Regulation Governance

5(1):90–117.

Canales R (2014) Weaving straw into gold: Managing organi-

zational tensions between standardization and ﬂexibility in

microﬁnance. Organ. Sci. 25(1):1–28.

Coslovsky S (2011) Relational regulation in the Brazilian Ministério

Publico: The organizational basis of regulatory responsiveness.

Regulation Governance 5(1):70–89.

Daley-Harris S (2009) State of the microcredit summit campaign

report 2009. Report, Microcredit Summit Campaign, Wash-

ington, DC. http://www.microcreditsummit.org/resource/48/

state-of-the-microcredit-summit.html.

Durkheim E (1997) The Division of Labor in Society (Free Press,

New York).

Emerson RM (1962) Power-dependence relations. Amer. Sociol. Rev.

27(1):31–41.

Fernandez RM, Greenberg J (2013) Race, network hiring, and sta-

tistical discrimination. McDonald S, ed. Networks, Work and

Inequality, Research in the Sociology of Work, Vol. 24 (Emerald

Group Publishing, Bingley, UK), 81–102.

Field E, Pande R (2008) Repayment frequency and default in

microﬁnance: Evidence from India. J. Eur. Econom. Assoc. 2(3):

501–509.

Fishman R, Paravasini D, Vig V (2011) Social proximity and loan

outcomes: Evidence from an Indian bank. Working paper,

Columbia University, New York.

Fudenberg D, Tirole J (1990) Moral hazard and renegotiation in

agency contracts. Econometrica 58(6):1279–1319.

Gibbons R, Henderson R (2011) Relational contracts and organiza-

tional capabilities. Organ. Sci. 23(5):1350–1364.

Giné X, Karlan D (2012) Group versus individual liability: Long

term evidence from Philippine microcredit lending groups.

Working paper, World Bank.

Giné X, Jakiela P, Karlan D, Morduch J (2010) Microﬁnance games.

Amer. Econom. J.: Appl. Econom. 2(3):60–95.

Granovetter M (1992) Problems of explanation in economic sociol-

ogy. Nohria N, Eccles R, eds. Networks and Organizations: Struc-

ture, Form and Action (Harvard Business School Press, Boston),

25–56.

Granovetter MS (1973) The strength of weak ties. Amer. J. Sociol.

78(6):1360–1380.

Gulati R (1998) Alliances and networks. Strategic Management J.

19(4):293–317.

Guseva A, Rona-Tas A (2001) Uncertainty, risk, and trust: Russian

and American credit card markets compared. Amer. Sociol. Rev.

66(5):623–646.

Hannan M, Burton M, Baron J (1996) Inertia and change in the early

years: Employment relations in young, high-technology ﬁrms.

Indust. Corporate Change 5(2):503–536.

Hart O, Moore J (1999) Foundations of incomplete contracts. Rev.

Econom. Stud. 66(1):115–138.

Heimer M (1992) Doing your job and helping your friends: Uni-

versalistic norms about obligations to particular others in net-

works. Nohria N, Eccles R, eds. Networks and Organizations:

Structure, Form and Action (Harvard Business School Press,

Boston), 143–164.

Hermalin B, Katz AW, Craswell R (2007) Contract law. Polinsky A,

Shavell S, eds. Handbook of Law and Economics, Vol. 1 (North-

Holland, Amsterdam), 3–138.

Hertzberg A, Liberti J, Paravisini D (2010) Information and incen-

tives inside the ﬁrm: Evidence from loan ofﬁcer rotation.

J. Finance 65(3):795–828.

Heskett J, Hallowell R (1993) Southwest Airlines—1993 (A). HBS

Case 694-023, Harvard Business School, Boston.

Jain P, Moore M (2003) What makes microcredit programmes effec-

tive? Fashionable fallacies and workable realities. Working

paper, Institute of Development Studies, Brighton, UK.

Janik M (2012) How is high staff turnover impacting Mexico’s

market? Announcement (December 27), Microﬁnance Informa-

tion Exchange, Washington, DC. http://www.mixmarket.org/

announcements/How-is-high-staff-turnover-impacting-Mexico

%E2%80%99s-market%3F%20%20.

Lipsky M (1980) Street Level Bureaucracy: Dilemmas of the Individual

in Public Services (Russell Sage Foundation, New York).

Macaulay S (1963) Non-contractual relations in business: A prelim-

inary study. Amer. Sociol. Rev. 28(1):55–67.

Macneil I (1978) Contracts: Adjustment of long-term economic rela-

tions under classical, neoclassical, and relational contract law.

Northwestern Univ. Law Rev. 72(6):854–902.

Malhotra D, Murnighan J (2002) The effects of contracts on inter-

personal trust. Admin. Sci. Quart. 47(3):534–559.

May P, Winter S (2000) Reconsidering style of regulatory enforce-

ment: Patterns in Danish agro-environmental inspection. Law

Policy 22(2):143–173.

McMillan J, Woodruff C (1999) Interﬁrm relationships and informal

credit in Vietnam. Quart. J. Econom. 114(4):1285–1320.

Moran P (2005) Structural vs. relational embeddedness: Social

capital and managerial performance. Strategic Management J.

26(12):1129–1151.

Morduch J (1994) Poverty and vulnerability. Amer. Econom. Rev.

84(2):221–225.

Morduch J, Armendáriz de Aghion B (2005) The Economics of Micro-

ﬁnance (MIT Press, Cambridge, MA).

Nahapiet J, Ghoshal S (1998) Social capital, intellectual capital,

and the organizational advantage. Acad. Management Rev. 23(2):

242–266.

Petersen M, Rajan R (1994) The beneﬁts of lending relationships:

Evidence from small business data. J. Finance 49(1):3–37.

Piore M (2011) Beyond markets: Sociology, street-level bureaucracy,

and the management of the public sector. Regulation Governance

5(1):145–164.

Canales and Greenberg: Loan Ofﬁcer Consistency and Exchange Continuity

Poppo L, Zenger T (2002) Do formal contracts and relational gover-

nance function as substitutes or complements? Strategic Man-

agement J. 23(8):707–725.

Reille X, Forster S (2008) Foreign capital investment in microﬁ-

nance: Balancing social and ﬁnancial returns. Focus Note 44,

CGAP, Washington, DC.

Rousseau DM (1990) New hire perceptions of their own and their

employer’s obligations: A study of psychological contracts.

J. Organ. Behav. 11(5):389–400.

Sanchez-Burks J, Nisbett R, Ybarra O (2000) Cultural styles, rela-

tional schemas, and prejudice against out-groups. J. Personality

Soc. Psych. 79(2):174–189.

Sanyal P (2009) From credit to collective action: The role of micro-

ﬁnance in promoting women’s social capital and normative

inﬂuence. Amer. Sociol. Rev. 74(4):529–550.

Schelling T (1960) The Strategy of Conﬂict (Harvard University Press,

Cambridge, MA).

Seabright MA, Levinthal DA, Fichman M (1992) Role of individual

attachments in the dissolution of interorganizational relation-

ships. Acad. Management J. 35(1):122–160.

Sharpe S (1990) Asymmetric information, bank lending, and

implicit contracts: A stylized model of customer relationships.

J. Finance 45(4):1069–1087.

Smith ER (1984) Model of social inference processes. Psych. Rev.

91(3):392–413.

Sorenson O, Rogan M (2014) (When) do organizations have social

capital? Annual Rev. Sociol. 40:261–280.

Stiglitz J (1990) Peer monitoring and credit markets. World Bank

Econom. Rev. 4(3):351–366.

Stiglitz J, Weiss A (1986) Credit rationing and collateral. Edwards J,

Franks J, Mayer C, Schaefer S, eds. Recent Developments in Cor-

porate Finance (Cambridge University Press, Cambridge, UK),

101–135.

Uzzi B (1997) Social structure and competition in interﬁrm net-

works: The paradox of embeddedness. Admin. Sci. Quart. 42(1):

35–67.

Uzzi B, Lancaster R (2003) Relational embeddedness and learning:

The case of bank loan managers and their clients. Management

Sci. 49(4):383–399.

Van Maanen J (1973) Observations on the making of policeman.

Human Organ. 32(4):407–418.

Van Maanen J (1978) The asshole. Manning PK, Van Maanen J, eds.

Policing: A View from the Streets (Goodyear, Santa Monica, CA),

231–238.

Van Maanen J (1991) The smile factory: Work at Disneyland. Frost

PJ, Moore LF, Louis MR, Lundberg CC, Martin J, eds. Reframing

Organizational Culture (Sage, Newbury Park, CA), 58–76.

Weber M (1962) Basic Concepts in Sociology (Citadel Press,

New York).

Weber M (1978) Economy and Society (University of California Press,

Berkeley).

Williamson O (1985) The Economic Institutions of Capitalism (Free

Press, New York).

Yunus M (2003) Banker to the Poor: Microlending and the Battle Against

World Poverty (Public Affairs, New York).

Zucker L (1986) Production of trust: Institutional sources of eco-

nomic structure, 1840–1920. Staw BW, Cummings LL, eds.

Research in Organizational Behavior, Vol. 9 (JAI Press, Greenwich,

CT), 53–111.

Community influence on microfinance loan defaults under crisis conditions: Evidence from Indian demonetization

Article

Oct 2023
STRATEGIC MANAGE J

Research Summary The microfinance “group lending” approach has achieved widespread success in promoting high rates of repayment, and thus the viability of financial access, in very low‐income environments. Yet group lending, which relies on social connections between borrowers to reinforce repayment, may be vulnerable under crisis conditions in which defaults are commonplace. We explore this possibility in the context of the liquidity crisis that followed India's 2016 demonetization policy. Using proprietary data on the repayment behavior of about two million microfinance borrowers, we find evidence of disproportionate localization of defaults within lending communities. Further analysis reveals evidence consistent with borrower‐to‐borrower spread of defaults not only through formal joint‐liability connections but also through informal community‐level connections, the latter effect being especially pronounced for borrowers from the same religion. Managerial Summary Microfinance lenders have successfully employed a “group lending” approach that holds borrowers within a small group responsible for each other's loans, thus creating strong social pressures for repayment. The findings of this study underscore potential vulnerabilities in the group lending model during economic crises. We analyze the loan repayment behavior of two million microfinance borrowers during a liquidity crisis precipitated by India's 2016 demonetization policy, finding that the resulting defaults were clustered within particular lending communities. Further analysis suggests that social processes within communities played a role in spreading defaults, not only through formal ties between borrowers who were responsible for each other's repayments, but also through informal social ties. The estimated effect of informal social ties was particularly strong for borrowers who shared the same religion.

You Say Social Agenda, I Say My Job: Navigating Moral Ambiguities by Frontline Workers in a Social Enterprise

Article

Full-text available

Sep 2023
J BUS ETHICS

Building on the emerging literature on the ethics of social enterprises (SEs), this paper advances the underexplored role of frontline workers (FLWs) as embedded agents at the interface between communities and SEs. Specifically, we uncover the subjectivity of FLWs as they navigate moral ambiguities while performing their professional roles, dealing with rules and regulations within the organizational hierarchy and living as members of local communities. Based on an inductive case study of a microfinance organization in Cameroon, we find that FLWs engage in three rationalization strategies: cautious disengaging, safeguarding self-interest, and justifying relevance. Our findings offer a better understanding of the ethics of SEs by unpacking the subjectivity of FLWs. We highlight a bottom-up account of caring SEs, identify a boundary condition to subjectivity, and present a nuanced view of FLWs in embedded organizations. We also discuss the practical implications for SEs to improve their compassion and cater for the mental wellbeing of FLWs.

When Do Haters Act? Peer Evaluation, Negative Relationships, and Brokerage

Article

Full-text available

Jan 2024

Evidence in Practice: How Structural and Programmatic Scaffolds Enable Collaboration in International Development

Article

Full-text available

Apr 2024

This inductive study of eight international development interventions analyzes mechanisms that enable integration of evidence in practice, a perennial challenge of learning and collaboration across occupational and organizational boundaries. We demonstrate how structural and programmatic scaffolding practices enabled actors from an array of organizations and communities of practice to collaborate and learn despite the uncertainty and complexity inherent in the international development context. These modular scaffolding practices offered temporary stabilization and support that fostered the counter-normative behaviors and mindsets required for continuous learning and adaptive coordination. Through 226 in-depth interviews with international development experts, including practitioners in eight matched interventions in India, Mexico, South Africa, and Ghana, we identified and analyzed mechanisms that explain the varying effectiveness with which evidence was integrated in each case. Our findings have implications for interorganizational innovation and collaboration under conditions of complexity and uncertainty and for dynamic interactions among individuals, their organizations, and their communities of practice when they are attempting to bring about systemic change.

Stepping stone: the logic of financial inclusion through microfinance – evidence from China’s largest commercial microfinance institution

Article

Feb 2025

Purpose This study assesses the impact of microfinance experience on rural households’ subsequent bank credit access, defined as the stepping-stone effect. We evaluate the average impact, heterogeneity and mechanisms of this effect and draw policy implications. Design/methodology/approach Based on 2018 survey data from 2,347 rural households in six Chinese provinces, this study uses the IV-Probit model to evaluate the stepping-stone effect. Findings The microfinance experience increases rural households’ probability of obtaining subsequent bank credit by six percentage points. Whether distinguished by social capital or economic wealth, this effect is at least as significant among households previously excluded from bank credit as it is among those originally favored by bank credit. Two mechanisms behind this effect are identified: (1) the microfinance experience provides financial education, transforming hidden demand for bank credit into effective demand and (2) the microfinance experience creates a credit record in the National Credit Information System, increasing the ability to obtain bank credit by providing banks with credible credit information. Originality/value This study provides evidence of a progressive relationship in which microfinance acts as a stepping stone to deliver “credit graduate” clients to bank credit. And we also provide a different perspective for understanding MFIs’ relatively high interest rates. We may treat a part of the interest payment as a tuition for financial education and a price for credit information. Highlights (1) We study whether microfinance in China had the stepping-stone effect of improving a rural household’s subsequent access to bank credit and investigate the likely mechanisms behind it. (2) Microfinance served as a stepping stone to bank credit: the microfinance experience increases rural households’ probability of obtaining subsequent bank credit by six percentage points. (3) Whether distinguished by social capital or economic wealth, this effect is at least as significant among households previously excluded from bank credit as it is among those originally favored by bank credit. (4) On the credit demand side, the microfinance experience provides financial education, transforming hidden demand for bank credit into effective demand. (5) On the credit supply side, the microfinance experience creates a credit record in the National Credit Information System, increasing the ability to obtain bank credit by providing banks with credible credit information.

Referral Triads

Article

Dec 2024
ADMIN SCI QUART

Third parties who refer clients to expert service providers help clients navigate market uncertainty by curating well-tailored matches between clients and experts and by facilitating post-match trust. We argue that these two functions often entail trade-offs because they require referrers to activate network relationships with different experts. While strong ties between referrers and experts promote trust between clients and experts, the presence of such ties reduces the likelihood that intermediaries refer clients to socially distal experts who may be better suited to serve clients’ needs. We examine this central and unexplored tension by using full population medical claims data for the state of Massachusetts. We find that when primary care physicians (PCPs) refer patients to specialists with whom the PCPs have strong ties, patients demonstrate more confidence in the specialists’ recommendations. However, a strong tie between the PCP and specialist also reduces the expertise match between a patient’s health condition and a specialist’s clinical experience. These findings suggest that the two central means by which referrers add value may be at odds with one another because they are maximized by the activation of different network ties.

How business model innovation facilitates microcredit in balancing social mission with commercial performance - evidence from local commercial banks

Article

May 2024
TECHNOL FORECAST SOC

Strategies to Improve the Sales Performance of Car Loan Products in Bank Shinhan Indonesia

Article

Mar 2024

In internal Bank Shinhan, My Car Loan is one of their top consumer products. However, when compared to other car financing competitors, Bank Shinhan was far behind in terms of sales performance. Hyundai Motors Indonesia has four official financial partners to support its car purchasing, namely: Maybank Finance, Mandiri Tunas Finance, Sunindo KB Finance, and Bank Shinhan. The company is the only bank-based financial partner, while the three others were multi-finance companies that specialized in car financing. Moreover, two of them, Mandiri Tunas Finance and Maybank were already well-known players in the car financing industry. Meanwhile, Bank Shinhan is a new player with low brand awareness among the public. In this research, the data collection methods were interviews and questionnaires. As for the research method, for the questionnaire data, the author used the NPS (Net Promoter Score) method and reliability test. As for the interview data, the author used the TA (thematic analysis) method. With this research, the author aimed to get insights to propose new strategies.

Microcredit, Gender and Geography. Case of Banecuador

Chapter

Jan 2024

This article investigates the factors to access to credit of the largest public bank in Ecuador. The methodology includes a logistic regression model to a database corresponding to loan applications from associations of the Popular and Solidarity Economy of Ecuador during the period 2016–2019. The estimates show that the unsecured guarantee, belonging to the tertiary sector, being located in a rural environment and the sex of the loan officer influence the access to financing. In the first place, it was found that the businesses that belong to the tertiary sector show a higher statistically significant probability for the granting of credit, this is in accordance with the business structure of Ecuador where most of the companies are part of this sector.

Loan Officer Specialization and Credit Defaults

Article

Dec 2023
J BANK FINANC

Credit rationing and collateral

Chapter

Sep 1986

This book was originally published in 1986. During the decade preceding publication there were a number of significant developments in financial economics and major contributions made both by individuals who could be classified as conventional financial economists and by others who do not fit easily into this category - theoretical microeconomists, public and industrial economists. This volume contains a selection from the papers presented at a conference in Oxford in September 1985 which aimed to bring together a number of the leading participants in this field. The papers in the volume cover a wide range of topics - the efficiency of financial markets, new equity issues, asymmetric corporate taxation and investment, credit rationing, international investment, the foundations of banking theory - but they clearly reflect the main themes in financial economics at the time: the importance of informational asymmetries and of taxation.

Role of Individual Attachments in the Dissolution of Interorganizational Relationships

Article

Mar 1992

Information and Incentives Inside the Firm: Evidence from Loan Officer Rotation

Article

Jan 2009

Group Lending, Repayment Incentives, & Social Collateral

Article

Jan 1995
J DEV ECON

Networks and alliances

Article

Apr 1998
STRATEGIC MANAGE J

Ranjay Gulati

This paper introduces a social network perspective to the study of strategic alliances. It extends prior research, which has primarily considered alliances as dyadic exchanges and paid less attention to the fact that key precursors, processes, and outcomes associated with alliances can be defined and shaped in important ways by the social networks within which most firms are embedded. It identifies five key issues for the study of alliances: (1) the formation of alliances, (2) the choice of governance structure, (3) the dynamic evolution of alliances, (4) the performance of alliances, and (5) the performance consequences for firms entering alliances. For each of these issues, this paper outlines some of the current research and debates at the firm and dyad level and then discusses some of the new and important insights that result from introducing a network perspective. It highlights current network research on alliances and suggests an agenda for future research.© 1998 John Wiley & Sons, Ltd.

Microfinance: Where do we Stand?

Chapter

Jan 2004

Economies are built upon people buying and selling, lending and borrowing. The beauty of the market is that, when it works well, sellers are matched to buyers and lenders are matched to worthy borrowers. But when the market does not work well, goods go unsold and promising investment projects go unfunded. We understand why markets fail - the economics of information provides rigorous underpinnings for why credit markets, in particular, are so problematic.1 The challenge has been to move from diagnosis to prescription. The challenge is particularly great in poorer regions, where individuals may have workable ideas and relevant experience but lack collateral. Even a £100 loan can make a difference to a small-scale shopkeeper or craftsperson in countries like Nepal or Uganda, but formal sector banks have steered clear, focusing instead on larger loans to better-established, wealthier clients.

J. Social capital, intellectual capital, and the organizational advantage

Article

Jan 1998

Scholars of the theory of the firm have begun to emphasize the sources and conditions of what has been described as “the organizational advantage,” rather than focus on the causes and consequences of market failure. Typically, researchers see such organizational advantage as accruing from the particular capabilities organizations have for creating and sharing knowledge. In this article we seek to contribute to this body of work by developing the following arguments: (1) social capital facilitates the creation of new intellectual capital; (2) organizations, as institutional settings, are conducive to the development of high levels of social capital; and (3) it is because of their more dense social capital that firms, within certain limits, have an advantage over markets in creating and sharing intellectual capital. We present a model that incorporates this overall argument in the form of a series of hypothesized relationships between different dimensions of social capital and the main mechanisms and processes necessary for the creation of intellectual capital.

Google's quest to build a better boss

Article

Jan 2011
New York Times

A. Bryant

Race, Network Hiring, and Statistical Discrimination

Chapter

Jan 2013

Abstract: Purpose – Research has shown that employers often disfavor racial minorities - particularly African Americans - even when whites and minorities present comparable resumes when applying for jobs. Extant studies have been hard pressed to distinguish between taste-based discrimination where employers' racial animus is the key motivation for their poor treatment of minorities and variants of statistical discrimination where there is no assumption at all of racial animus on the part of the employer. This chapter proposes a test of these theories by observing whether employers use employee referrals as a “cheap” source of information to help assess applicant quality. Methodology/approach – Unique quantitative data encompassing the entire pool of 987 candidates interviewed by one company in the western United States during a 13-month period are used to test our arguments. Findings – We find that employers in this setting are making use of the cheap information available to them: Consistent with statistical discrimination theory, minority referrals are more likely to receive a job offer than non-referred minority applicants, and are not disfavored relative to referred whites. Originality/value of the chapter – Both statistical and taste-based theories of discrimination propose similar observable outcomes (lower rates of disfavored minority hiring). While different mental processes are being invoked by taste-based and statistical discrimination theories, the theories are extremely difficult to distinguish in terms of observable behaviors. Especially for the purpose of designing legal remedies and labor market policies to ameliorate the disparate treatment of minority groups, differentiating between these theories is a high priority.

Poverty and Vulnerability

Article

Jan 1994

Jonathan Morduch

A definition of stochastic poverty is set out in an attempt to formalise the difference between it and risk per se. Three additional factors contribute to poverty in poor countries: weather and price variability which are responsible for a large part of income fluctuations; financial institutions are poorly developed; and social insurance institutions are often weak. The problem of transitory poverty in rural India, much of which is due to stochastic poverty, is then addressed largely by increasing the extent and effectiveness of insurance mechanisms. -from Author