Pricing and Marketing Household Financial

Services in Developing Countries

Dean Karlan and Jonathan Zinman

Retail financial institutions worldwide

are facing greater competition and regu
latory scrutiny. This makes it increasingly
im portant for them to understand the driv
ers o f consumer demand for basic finan
cial services if they are to maximize prof
its, improve social impacts, and address
public policy concerns. Researchers also
need to understand these drivers in order
to calibrate, shape, and test models in fields
ranging from contract theory to behav
ioral economics to macroeconomics to basic
microeconomics. Likewise, policymakers
need to understand these drivers in order
to sift through a plethora o f potentially rel
evant theories and set appropriate regula
tions. Much o f our research seeks to iden
tify the effects o f pricing and marketing on
demand for short-term loan and savings
products in developing countries.
Pinning down causal effects o f financial
institutions pricing and marketing strategies
is complicated by at least five issues. One is
the classic social science problem: Relying
on observational data is fraught with the
risk that changes in price or marketing
are correlated with other changes in firm
strategy, in the macroeconomy, in house
hold budget constraints that drive selec
tion. This is a particular concern when esti
mating treatment effects from expanding
access to financial products such as credit,
savings, or insurance. A second issue, inti
mately related to the first, is low statistical
power due to limited variation in key policy
parameters. A firm making a single change
to pricing, a product, or marketing is basi
cally generating a single data point o f vari
ation. The effects o f the single change are
difficult to disentangle from other contem
poraneous changes affecting the firm and
its constituents. This is a particular concern
for savings products, as compared to loans,
since one-size-fits-all pricing is more com
mon and direct marketing is less common

with savings products. These two issues

are the primary motivation for employing
experimental methods.
A third complicating issue is that most
measures o f demand sensitivity for exam
ple, demand elasticities are not funda
mental or unchanging parameters. We
expect demand sensitivities to change with
factors like competition, labor market con
ditions, and search costs. A fourth issue is
that a firms levers are rarely perfect repre
sentations o f a single parameter. For exam
ple, variations in price, in particular, may
be confounded by other factors changing
simultaneously and may therefore lead to
deceiving results if interpreted strictly as an
estimate o f demand sensitivity. A fifth issue
is that strategy often requires an under
standing of underlying mechanisms, while
identifying mechanisms requires observ
ing off-equilibrium behavior. For example,
observing loan repayment and other bor
rower behaviors under atypical conditions
can help test theories o f asymmetric infor
mation or liquidity constraints.
We address these challenges using field
experiments implemented by financial insti
tutions in the course o f their day-to-day
operations. The partnering financial institu
tions randomly assign prices, communica
tions, or access to products, generating vari
ation that is uncorrelated with other factors
that vary endogenously over time or people.
This addresses Issue One above. The finan
cial institutions randomize policies at the
individual or neighborhood level in order to
generate sufficient statistical power to iden
tify causal effects. This addresses Issue Two.
In some instances, the financial institutions
randomized policies are implemented across
sufficiently different people or markets, and
are in place for long enough or with vary
ing lengths o f time, that we can examine
under what conditions demand varies. This
addresses Issue Three.1 In another instance,

Dean Karlan is a professor

o f economics at Yale University
and president o f Innovations for
Poverty Action, a nonprofit dedi
cated to discovering and promot
ing solutions to global poverty
problems. He is on the Executive
Committee o f the Board of
Directors and program chair for
finance o f the M IT Jameel Poverty
Action Lab. At Innovations for
Poverty Action, he is academic
co-chair of the Global Financial
Inclusion Initiative, the Small and
M edium Enterprise Initiative,
and the Ultra-Poor Safety Net
Graduation Initiative.
As a social entrepreneur,
Karlan is co-founder o f stickK.
com, a website that uses lessons
from behavioral economics to help
people reach personal goals. Karlan
received a Presidential Early Career
Award for Scientists and Engineers
in 2007, and was named an Alfred
P. Sloan Fellow the following
year. In 2011 he co-authored More
Than Good Intentions: How
a New Economics is Helping to
Solve Global Poverty. His educa
tion includes a Ph.D. in econom
ics from MIT, an M.B.A. and an
M.P.P. from the University of
Chicago, and a B.A. in interna
tional affairs from the University
ofVirginia.

N B E R Reporter *2015 Number 1

17

Jonathan Zinm an is a
professor o f economics at
D artm outh College, and co
founder and scientific direc
tor o f the U.S. Household
Finance Initiative (U SH FI)
of Innovations for Poverty
Action.
Zinmans research focuses
on household finance and
behavioral economics. He has
published papers in several top
journals in economics, finance,
and general-interest science,
and his work has been featured
extensively in popular and trade
media as well.
Z inm an
applies
his
research by working with pol
icymakers and practitioners
around the globe. He cur
rently serves on the inaugu
ral Consumer Advisory Board
o f the Consum er Financial
Protection Bureau, as a visiting
scholar at the Federal Reserve
Bank of Philadelphia, and as
a C om m unity Development
Research Advisory Council
member for the Federal Reserve
Bank of Boston. He also works
directly with financial service
providers, ranging from start
ups to nonprofits to publiclytraded companies, to develop
and test innovations that are
beneficial to both providers
and their clients.
18

N B E R Reporter 2015 Number 1

additional information that lenders could sensitivity; more recently, we worked with a
we use variation in price and advertising con produced evidence of some important impacts,
use to screen or price ex-ante and thereby large microlender in Mexico to randomize
tent to explore how those two levers interact, tested typically with some aspect of commit
interest rates at the level of 80 geographi
mitigate adverse selection.
addressing Issue Four.1
2 Lastly, through two- ment to the product,10 with several studies
cal regions across the country, with experi
A
fourth
finding
is
that
household
stage experimental designs, we have tackled pursuing further work to unpack mechanisms
mental rates in place for 30 months.16 This
demand
for
commitment
savings
balances
typically unobserved behavior on loan repay underlying the impacts.11
design allows us to estimate elasticities over
is
not
sensitive
to
price,
at
least
within
A third finding from our work is that
ment as well as returns to capital, which per
different time horizons that internalize any
the
range
of
market
rates
found
in
the
information asymmetries complicate lend
tains to Issue Five.3
spillovers (e.g., information transmission)
v
Philippines.14
This
is
somewhat
puzzling
In our work,4 and in the work of oth ers pricing strategies. Our work in the South
within
regions. We find that loan demand
in
light
of
our
next
set
of
findings

sub
ers,5 we learn that financial markets for credit Africa cash loan market and an individualis
more
or less unit elastic (-1) in year one,
stantial
price
sensitivity
to
consumer
credit
are not meeting the needs of the poor. In liability microloan market in the Philippines
with
price
sensitivity increasing over time to

interest
rates

although
we
emphasize
that
Mexico, the Philippines, and South Africa, finds evidence of substantial moral hazard.12
around
-3
in
year three. This degree of price
whether
this
finding
applies
to
other
types
we have found that financial institutions are These papers also suggest that this problem
sensitivity
is much larger than
able to expand access to microcredit
anything
else
found in the lit
by experimenting with risk-based pric
erature
to
date,
with the excep
ing models or building offices in new
Q U E S T IO N S A B O U T F IN A N C IA L SER VIC ES A R O U N D T H E GLOBE
tion
of
our
finding
on price
geographic areas, effectively reducing
increases
in
South
Africa.17
We
the price of financial institution credit
attribute
this
to
our
designs
from infinity to a market rate for cer
ability to capture a long-run
tain borrowers. Others have found the
equilibrium, as opposed, for
same to be true in Morocco, Bosnia
example, to a temporary and
and Herzegovina, Mongolia, India,
isolated promotion.
and Ethiopia. Indeed, every study of
But our most surpris
which we are aware that has examined
ing
finding on price sensitiv
the impacts of expanding credit supply
ity
comes
from a new paper
has found that the expansion increased
on
a
large
Turkish banks
borrowing and did not merely crowd
experiment
with
direct-mar
out other lenders. Similarly, finan
keting
of
an
overdraft
line of
cial institutions offering new-comcredit.18
Messages
mentioning
mitment savings products for 6-to-24
the cost of overdrafting reduce
month savings goals have found takeoverdraft
usage, even though
up rates typically around 20-30 per
those
messages
offer a 50 per
W h a t is t h e im p a c t
cent.6 Other financial institutions have
cent
rebate
on
overdraft
inter
o f e x p a n d e d access
found similar unmet demand for com
est:
Substantially
reducing
the
t o g r o u p le n d in g ?
mitment savings.7
price
of
the
commodity
reduces
W h a t a r e t h e e ff e c t s o f a d v e r t is in g
A second finding from the stud
demand for it. This finding
c o n t e n t o n r e a l d e c is io n s ?
ies above is that the marginal consum
is consistent with models of
ers of basic financial services derive a
S o u rc e s: N B E R W o r k in g P a p e rs 1 7 5 9 2 ,1 9 1 1 9 , 2 0 9 5 6 a n d Quarterly Journal o f Economics (2010)
shrouded equilibria in which
variety of financial benefits from them.
firms lack incentives to draw
This is an important reality check,
attention to add-on prices;19
given concerns that various biases in
this
and
other
findings in the paper, dis
of
savings
instruments
and
settings
is
an
can
be
addressed
with
stronger
dynamic
incen
household decision-making can lead to coun
cussed
below,
are
consistent with a model of
open
question.
tives
and
repayment
reminders
from
loan
offi
terproductive borrowing.8 Beyond the basic
limited
attention
and memory.20
A
fifth
finding
from
our
work
is
that
cers.
Another
paper
develops
an
experimen
reality check, the several studies that follow
A
sixth
finding
is that communications
*
household
demand
for
consumer
credit
is
tal
design
to
test
for
an
interaction
between
random assignment to loan or savings prod
of
various
types
can
greatly affect demand.
price
sensitive,
and
sometimes
in
surprising
ex-ante
and
ex-post
asymmetric
information
uct availability with extensive household and
Our
direct-marketing
experiment in South
ways.
Our
early
work
in
this
area
consisted
problems

in
our
setting,
selection
on
mal
microenterprise surveys have yielded surpris
Africa
randomized
mailer
content along
of
direct-mail
experiments
with
a
small-dolleability
to
repayment
incentives

and
does
ing findings. On the credit side, the results
side
price
and
found
the
advertising
content
lar
lender
in
South
Africa.15
Potential
bor
not
rule
out
an
empirically
important
interac
have yielded little support for microcredits
had
large
effects.21
We
find
some
evidence
rowers
were
price
sensitive,
but
not
elastic,
tion,
although
the
estimates
are
imprecise.13
great promise of poverty alleviation and social
with respect to price cuts [0 > elasticity > that content designed to trigger automatic
transformation. Rather, the benefits have been That paper also tests a remedy incentivized
-1] and were extremely elastic with respect responses was more effective than content
peer
referrals

and
finds
evidence
that
refer
modest, and concentrated more in household
to price increases. Direct-mail promotional designed to trigger deliberative responses,
ring
peers
are
very
helpful
in
pressuring
friends
risk management and flexibility than in profit
experiments have the drawback of identify but overall it was difficult to predict exactly
to
repay
ex-post
(thereby
mitigating
moral
haz
able microenterprise growth.9 On the savings
ing short-run rather than steady-state price which types of ad content would affect
ard).
It
does
not
find
evidence
that
peers
have
side, the first wave of impact evaluations has

demand based on prior work on behavioral

economics. Our experiment in Turkey var
ied messaging content and intensity as well
as overdraft pricing, and we found evidence
that both of these levers mattered greatly. In
contrast to the core finding on price that
mentioning it reduces demand for over
drafts simply mentioning overdraft avail
ability substantially increases demand. And
more intense messaging sending the
same message more often amplifies both
the demand-increasing effects of advertis
ing overdraft availability and the demanddecreasing effects of advertising an over
draft price reduction. On the savings side,
we have found, across three different banks
in three different countries, that sending
reminders to new commitment savings
account customers increases commitment
attainment.22 Messages that mention both
savings goals and financial incentives are
particularly effective, while other content
variations such as gain versus loss framing
do not have significantly different effects.
This set of studies speaks to the importance
of limited and malleable consumer atten
tion to household finances.23
The findings reviewed here, taken
together with the work of many other
researchers, are initial steps toward unpack
ing the nature and implications of house
hold demand for financial services. Our
recent review articles highlight many
opportunities for future work.24

1 D. Karlan and J. Zinm an, Long-Run

Price Elasticities o f D em andfor Credit:
Evidencefrom a Countrywide Field
Experiment in Mexico,N B E R Working
Paper No. 19106, June 2013; and M.
Angelucci, D. Karlan, and J. Zinman,
Win Some Lose Some? Evidencefrom
a Randomized Microcredit Program
Placement Experiment by Compartamos
Banco,N B E R Working Paper No. 19119,
June 2013.
2 S. Alan, M. Cemalcilar, D. Karlan, J.
Zinm an, Unshrouding Effects on Demand
fo r a Costly Add-on: Evidencefrom Bank
Overdrafts in Turkey,N B E R Working
Paper No. 20956, February 2015.
3 See D. Karlan and J. Zinman,
Observing Unobservables: Identifying
N B E R Reporter *2015 Number 1

19

Information Asymmetries with a

Kinnan, The Miracle o f Microfmance?
Credit: Too M uch or Too L ittle (or Just
Consumer Credit Field Experiment,
Evidencefrom a Randomized Evaluationf R ig h t)!N B E R Working Paper No.
Econometrica, 77(6), 2009, pp. 1993N B E R Working Paper No. 18950, May
19682, November 2013, an d Journal
ZOO8; G. Bryan, D. Karlan, and J.
2013, and American Economic Journal:
o f Legal Studies, 43 (S2) (Special Issue
Zinm an, Referrals: Peer Screening and
Applied Economics, 7(1), 2015, pp.22on Benefit-Cost Analysis o f Financial
Enforcement in a Consumer Credit Field
53; and B. Crepon, F. Devoto, E. Duflo,
Regulation), 2014, pp.S209-37.
Experiment,N B E R Working Paper No.
W Pariente, Estimating the Impact o f
9 See endnotes 4 and 5 fo r the relevant
17883, March 2012, and American
Microcredit on Those Who Take It Up:
citations.
Economic Journal: Microeconomics,
Evidencefrom a Randomized Experiment
10 See D. Karlan, A. Ratan, a n d J.
forthcoming; and L. Beaman, D. Karlan,
in Morocco," N B E R Working Paper
Z inm an, Savings by a n d fo r the
B. Thuysbaert, C. Udry, S elf Selection
No. 20144, M ay 2014, and American
Poor: A Research Review a n d Agenda,
into Credit Markets: Evidencefrom
Economic Journal: Applied Economics,
Review o f Income and W ealth, 6 0 (1),
Agriculture in Mali,N B E R Working
7(1), 2015, pp. 123-50.
2014, p p.3 6 -7 8 ; and J. Jamison, D.
Paper No. 20387, August 2014.
6 See D. Karlan and J. Zinm an,
Karlan, and J. Z inm an, Financial
4 See D. Karlan and J. Zinm an,
Price and Control Elasticities o f
Education and Access to Savings
Expanding Credit Access: Using
D em andfor S a vingsY ale Working
Accounts: Complements or Substitutes?
Randomized Supply Decisions to Estimate
Paper, 2014. http://karlan.yale.edu/p/
Evidence fro m Ugandan Youth Clubs,
the ImpactsReview of Financial Studies,
SavingsElasticities_2014_01_v9.pdf;
N B E R Working Paper No. 20135, M ay
23(1), 2010, pp.433-64; D. Karlan
D. Karlan, M. McConnell, S.
2014.
and J. Zinm an, Microcredit in Theory
Mullainathan, and J. Zinm an, Getting
11 See L. Beaman, D. Karlan, and
and Practice: Using Randomized Credit
to the Top o f M ind: How Reminders
B. Thuysbaert, Savingfor a (not So)
Scoringfo r Impact EvaluationScience,
Increase Saving,N B E R Working Paper
R ainy Day: A R andom ized Evaluation
332(6035), 2011, pp. 1278-84; M.
No. 16205, July 2010, and Management
o f Savings Groups in M ali,N B E R
Angelucci, D. Karlan, andJ. Zinm an,
Science,forthcoming; and N. Ashraf, D.
Working Paper No. 20600, October
Microcredit Impacts: Evidencefrom
Karlan, and W. Yin, Tying Odysseus to
2014; and M . Callen, S. De Mel, C.
a Randomized Microcredit Program
the Mast: Evidence from a Commitment
McIntosh, C. Woodruff, W hat Are
Placement Experiment by Compartamos
Savings Product in the Philippines,
the Headwaters o f Formal Savings1
Banco,N B E R Working Paper No. 19827,
Quarterly Journal o f Economics,
Experim ental Evidence fro m Sri Lanka,
January 2014, and American Economic
121(2), 2006, pp.635-72.
N B E R Working Paper No. 20736,
Journal: Applied Economics, 7(1),
7 See P. Dupas and J. Robinson,
December 2014.
2015, pp. 151-82; and A. Banerjee, D.
Savings Constraints and Microenterprise 12 See D. Karlan and J. Zinm an.
Karlan, and J. Zinm an, Six Randomized Development: Evidence from a Field
Observing Unobservables: Identifying
Evaluations o f Microcredit: Introduction
Experim ent in Kenya,N B E R Working
Information Asymmetries with a
and Further Steps,American Economic
Paper No. 14693, January 2009, and
Consumer Credit Field Experim ent,
Journal: Applied Economics, 7(1), 2015,
American Economic Journal: Applied
Econometrica, 77(6), 2009, pp. 1 9 9 3 pp.1-21.
Economics, 5(1), 2013, pp.163-92; P.
ZOOS; and D. Karlan, M . Morten, and
5 See 0. Attanasio, B. Augsburg,
Dupas and J. Robinson, W hy D ont the
J. Z inm an,
A Personal Touch: Text
R. De Haas, E. Fitzsimons, and H.
Poor Save M orel Evidence from Health
Messaging fo r Loan Repayment,N B E R
Harmgart, The Impacts ofMicrofmance:
Savings Experiments,N B E R Working
Working Paper No. 17952, March 2012.
Evidencefrom Joint-Liability Lending
Paper No. 17255, July 2011, and
13 G. Bryan, D. Karlan, an d J.
in Mongolia,American Economic
American Economic Review, 103(4),
Z inm an, Referrals: Peer Screening and
Journal: Applied Economics, 7(1), 2015,
2013, pp. 1138-71; and S. Prina,
Enforcement in a Consumer Credit Field
pp.90-122; B. Augsburg, R. De Haas, H.
Banking the Poor via Savings Accounts:
Experim ent,N B E R Working Paper No.
Harmgart, and C. Meghir, The Impacts
Evidencefrom a Field E xperim ent,
17883, March 2012, and American
o f Microcredit: Evidencefrom Bosnia and
Journal o f Development Economics,
Econom ic Journal: M icroeconomics,
Herzegovina,N B E R Working Paper No.
forthcoming.
forthcoming.
18538, November 2012, and American
8 For reviews o f decision-making biases
14 D. K arlan a n d J. Z in m a n , Price
Economic Journal: Applied Economics,
and other potential sources o f inefficiency
a n d Control Elasticities o f D em and
7(1), 2015, pp. 183-203; A. Tarozzi, J.
in markets fo r household debt and savings fo r Savings, Yale Working Paper,
Desai, and K. Johnson, The Impacts o f
products, see D. Karlan, A. Ratan, and
2014. http://karlan.yale.edu/pZ
Microcredit: Evidencef o m Ethiopia
J. Zinm an, Savings by andfor the Poor:
SavingsElasticities_2014_ 01_v9.pdf.
American Economic Journal: Applied
A Research Review and Agenda,Review
15 D. K arlan a n d J. Z inm a n ,
Economics, 7(1), 2015, pp.54-89; E.
o f Income and Wealth, 60(1), 2014,
Credit Elasticities in LessDuflo, A. Banerjee, R. Glennerster, C. G.
pp.36-78; and J. Zinm an, Consumer
Developed Economies: Implications
20

N B E R Reporter 2015 Number 1

fo r Microfinance.The American
Competitive Markets, N B ER Working
Economic Review, 98(3), 2008,
Paper No. 11755, November 2005
1040-68.
and Quarterly Journal of Economics
16 D. Karlan and J. Zinman, Long121(2), 2006, 505-40.
Run Price Elasticities o f Demand for
20 P. Bordalo, N. Gennaioli, and A.
Credit: Evidence from a Countrywide
Shleifer, Memory, Attention, and
Field Experiment in Mexico,NBER
Choice,Royal Holloway University o f
Working Paper No. 19106, June 2013.
London February 2015 Working Paper,
17 For a review o f literatures on loan
http://scholar.harvard.edu/files/shleifer/
pricing and price sensitivity, as well
files/evokedsets_march2_ O.pdf.
as several o f the other literatures refer
21 M. Bertrand, D. Karlan, S.
enced here, see J. Zinman, Household
Mullainathan, E. Shafir and J.
Debt: Facts, Puzzles, Theories, and
Zinman, W hats Advertising Content
Policies,NBER Working Paper 20496,
Worth? Evidence from a Consumer
September 2014, and Annual Review of Credit Marketing Field Experiment,
Economics, 7, forthcoming.
Quarterly Journal of Economics, 12(1),
18 S. Alan, M. Cemalcilar, D. Karlan,
2010, p p .2 6 3 -3 0 5 .
and J. Zinman, Unshrouding Effects on 22 D. Karlan, M. McConnell, S.
D em andfor a Costly Add-on: Evidence
Mullainathan, and J. Zinman, Getting
from Bank Overdrafts in TurkeyJ
to the Top o f M ind: How Reminders
NBER Working Paper No. 20956,
Increase Saving,N B ER Working
February 2015.
Paper No. 16205, July 2010, and
19 See, e.g, X. Gabaix and D. Laibson,
Management Science,forthcoming.
Shrouded Attributes, Consumer
23 See also our work on priming (non-)
Myopia, and Information Suppression in effects on demand fo r (microcredit)

microinsurance: A. Zwane, J. Zinman,

E. Van Dusen, W. Pariente, C. Null, E.
Miguel, M. Kremer, D. S. Karlan, R.
Hornbeck, X. Gine, E. Duflo, F. Devoto,
B. Crepon, and A. Banerjee, Being
Surveyed Can Change Later Behavior
and Related Parameter Estimates,
Proceedings of the National Academy
of Sciences, 108(5), 2011, pp.1821-26;
and V. Stango and J. Zinman, Lim ited
and Varying Consumer Attention:
Evidencefrom Shocks to the Salience o f
Bank Overdraft Fees,N B ER Working
Paper No. 17028, May 2011, and
Review of Financial Studies, 27(4),
2014, pp. 990-1030.
24 See D. Karlan, A. Ratan, and J.
Zinman, Savings by andfor the Poor:
A Research Review and Agenda,Review
of Income and Wealth, 60(1), 2014,
pp.36-78; and J. Zinman, House
hold Debt: Facts, Puzzles, Theories, and
Policies,NBER Working Paper 20496,
September 2014, and Annual Review of
Economics, 7, forthcoming.

N B E R Reporter 2015 Number 1

21

Copyright of NBER Reporter is the property of National Bureau of Economic Research, Inc.
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.