What Are the Effects of Adding Credit and Insurance To a Conditional Cash Transfers Program in Mexico?

Manuela Angelucci, Carlos Chiapa and Silvia Prina 

Recent research suggests that chronic complexity and stress inherent in the lives of the poor can impede and hamper financial decision-making. Due to pressing short-term needs of the here and now, forward-looking behaviors, such as investing in education and health, may not get prioritized by the poor. Even with the knowledge and means, making changes in behavior (such as more time parenting or other healthy activities that could have far-reaching consequences) is challenging due to the lack of adequate resources.

Since a major source of anxiety in the life of the poor is related to financial complexity and uncertainty, for our project we assumed that simplifying the financial lives of the poor would create favorable circumstances for them to recognize the longer-term benefits of and willingly invest in their children’s education (through parenting and schooling) and adopt healthy behaviors. In addition, we anticipated that making the financial lives of the poor more straightforward would have a positive effect on their welfare.

Our focus group interviews with poor households in Mexico corroborated past research in showing that financial instability and emergencies are a major source of stress and anxiety. Therefore, for our target population we hypothesized that access to credit and insurance would improve the ability to cope with immediate shocks and that, in turn, these gains might promote the psychological well being of the poor and favor investing resources in education and health. We also had two more specific hypotheses: (1) recipients of the conditional cash transfer plus credit and insurance would have better parenting and higher adherence to healthy habits than recipients of only the basic conditional cash transfer; (2) experiencing unexpected income shocks reduces healthy habits and quality of parenting by adversely affecting psychological health, but this is less so for recipients of the conditional cash transfer plus credit and insurance.

We tested these hypotheses using as our target population the recipients of PROSPERA (formerly known as Oportunidades), Mexico’s flagship conditional cash transfer anti-poverty program, which covers approximately 25% of the Mexican population. This is a policy-relevant target population for two reasons. First, because conditional cash transfer programs targeting the poor are implemented in at least 33 countries worldwide and serve 90 million people in Latin America alone. Second, because the typical conditional cash transfer program requires its recipients to have regular health checks and provides them with health and nutrition education, both of them free of charge. Therefore, we target a population that has neither knowledge nor financial constraints to the adoption of health behaviors, but rather those whose mental wellbeing may be hurt by the stress and complexity of their lives.

The recipients, almost all women, received cash transfers in a bank account every other month. In addition, a subset of participants received additional benefits from a financial inclusion module called “Programa Integral de Inclusión Financiera ‘Prospera más con BANSEFI” (PROIIF), a supplemental program that provides beneficiaries with additional benefits at below-market prices. In particular, the main objective of PROIIF is to improve PROSPERA’s beneficiaries’ formal financial inclusion. Through PROIIF beneficiaries have access to the following products: an additional savings account (Ahorro Más con BANSEFI); two lines of credit at an interest rate substantially below market levels (Crédito Básico Más con BANSEFI and Crédito Más con Ahorro); a life insurance policy at a discounted price (Seguro Más con BANSEFI); and additional benefits (Paquete de Beneficios Adicionales BANSEFI).

To test hypothesis (1), we compared healthy habits, parenting, children’s schooling, mental health, cognition, stress and total household income of PROIIF and non-PROIIF households. The difference in outcomes for PROIIF and non-PROIIF recipients identified the average treatment effect of PROIIF under the assumptions that (i) PROIIF affects only its recipients and not other subjects (e.g., it rules out spillover effects of PROIIF) and (ii) there are no systematic differences between PROIIF and non-PROIIF recipients.

The first assumption is likely to hold because PROIIF and non-PROIIF recipients were geographically distant from each other. Therefore, spillover effects are unlikely. To check the validity of the second assumption, we compared the predetermined socio-economic characteristics of PROIIF and non-PROIIF recipients and their households. We find that these two groups of recipients and their households are fairly similar.

Figure 1. shows that healthy habits, parenting, children’s schooling, mental health, and cognition do not differ between PROIIF and non-PROIIF beneficiaries, thus rejecting our first hypothesis.

Figure 1.

To test hypothesis (2), we compared the healthy habits, parenting, children’s schooling, mental health, and cognition of households that have and have not suffered from unexpected shocks in the previous 14 days prior to the interview date. Since the experience of unexpected shocks may not be random, as poorer and more vulnerable households are more prone to suffer these shocks, we restricted the sample to households who had experienced at least one unexpected shock in the previous 12 months. Therefore, the variation that we exploited was in the timing, and not the experience, of these shocks. 

The difference in outcomes for households that were hit by an unexpected shock in the previous 14 days and households that experienced these shocks in the previous 12 months (but earlier than two weeks before the interview date), identified the effect of unexpected shocks under the assumptions that (i) the shocks affect only their recipients and not other subjects and (ii) there are no systematic differences between households that experienced the shocks sooner and later. 

Spillover effects of these shocks are unlikely, as only 7.16 percent of the sample experienced shocks in the previous 14 days prior to the interview date and these households were geographically spread out, so the data supports the first assumption. Moreover, the socio-economic characteristics of the two groups of households were similar, and consistent with the second hypothesis. 

Figure 2. shows that transitory and more permanent shocks reduced the income of PROSPERA’s beneficiary households by similar magnitude as the effects on income were not statistically different from each other. Permanent shocks did not consistently or overall significantly reduce adherence to healthy financial habits. Similarly, we see no effects on schooling or parenting, or on parental aspirations and expectations for their children’s schooling. 

Figure 2.

Further analysis shows that the magnitudes of the shocks of more permanent and transitory shocks on the outcomes of interest are not statistically different for PROIIF beneficiaries. There are two exceptions. First, receiving PROIIF attenuates the negative effects of experiencing transitory income shocks on household income. Second, while receiving PROIIF does not attenuate the negative effects of experiencing more permanent income shocks on income, it increases the amount of time desired to spend with children. 

Overall, findings from our data indicate that recipients of the conditional cash transfer plus credit and insurance do not seem to have better parenting and higher adherence to healthy habits than recipients of only the basic conditional cash transfer. In addition, while it is the case that experiencing unexpected transitory income shocks reduces income, the magnitude of the effect is lower for recipients of the conditional cash transfer plus credit and insurance. Finally, when experiencing more permanent income shocks, the quality of parenting by increases for recipients of the conditional cash transfer plus credit and insurance.

Read more in Angelucci, Chiapa and Prina's Final Report

Bank Backdoors, Burials, and Betting: Session Three of the 2016 Conference

The panel on "Accessories to Technology: Mobile Financial Services, Risk, and Insurance with discussant Ananya Roy of UCLA addressed how institutions must adjudicate claims and manage liability in volatile environments of new products, precarious populations, and financial experiments.  "Assessing the Need and Feasibility for Using Pre-Paid Card Technology in Delivering Added Services to Micro Finance Customers in Selected Regions of Uttar Pradesh: (India) by Debashis Acharya of the University of Hyderabad and Tapas Kumar Parida of the State Bank of India led off the session with a focus on the third largest state in India, where about 22 MFIs (microfinance institutions) operate.

New 2014 regulations have been reshaping the role of IRDA, the statutory body that regulates the insurance sector in India, which is tasked with both protecting the interests of policyholders while also ensuring the growth of the insurance industry.  Knowledge partners in the study included Bajaj Allianz Insurance, M2P Solutions (a prepaid card provider), and Utkarsh Micro finance, which is one of the leading 25 MFIs in India, according to a CRISIL 2014 report. They traced how a claim settlement process might evolve along three trajectories: 1) the Traditional/Conventional Model in which the MFI collects documents from clients after the death of the insured and submits materials to the insurance company, 2) the model of Electronic Transfer (NEFT) to the bank, which opens up possibilities for alternative payment mechanisms and split payment paradigms, and 3) the Open Loop Pre-Paid Card model in which the nominee gets directly benefitted by this process and in which unsettled claims can be reissued and fresh claims can  processed by pre-paid card.  Acharya described the costs and benefits of pre-paid cards from the perspective of users who think about mobile phones primarily as devices for communications.

"The Curious Case of Mobile Micro-insurance in South Africa: A View from Above and Below" (South Africa) by Christopher Paek of London School of Economics focused on Xhosa funerals and financial risk mitigation through insurance.  To demonstrate the importance of the issue of funeral insurance, he began with the case of Godfrey, who maintains a household composed of a wife, three children and a mother in a township outside Cape Town. With a monthly income of R2000 (about $130) it would be difficult to manage costs generated by the death of his father-in-law, which would include multiple undertakers, transportation to ancestral homeland, ceremonial slaughter of one cow (and a second cow for a male head-of-household), food for guests, and the slaughter of a sheep for the funeral banquet.  All tallied, such costs would be R41,780 or about  21 months of Godfrey's salary.  Rather than turn to the formal sector of conventional insurance, most planning for family funerals would depend on informal mechanisms, such as burial societies, family and friends, churches, or Mashonisas (loan sharks).  Funeral parlors themselves could serve as either formal and informal partners in contingency planning.

Paek explained his methodology of mixed qualitative and quantitative methods and his choices to integrate ethnographic methods in his work at the primary site in Khayelitsha, Cape Town, South Africa.  His data was collected from 23 formal sector interviews, which drew on informants in insurance companies, MNOs, and TSPs, as well as regulators/legislators, administrators, and industry representatives.  He also conducted 6 informal sector interviews with funeral parlors and burial societies, as well as client interviews with 76 survey respondents and 47 focus group respondents.

As inspiration Paek cited the work of Camilo Téllez and plugged his 2012 paper on "Emerging Practics in Mobile Microinsurance."  Now that telecommunications companies and mobile money firms were becoming interested in the funeral insurance market, there were even possibilities for paying for funeral coverage with airtime spending. Paek described M-insurance as "fertile ground" for innovative products and presented both a "view from above" and a "view from below" that was informed by Evans' and Pirchio's 2015 research oriented around an empirical examination of why mobile money schemes may flourish in one country and flounder in another. Like other IMTFI researchers he pointed to concepts and notions of trust.

He argued that mobile money might be slower to take off in the context of high crime rates, lack of access to formal legal recourse, inundation by scams, lack of consumer advocacy, saturation with fly-by-night operations, high unemployment rates (which erodes trust in social networks), and a proliferation of scams on the phone.  All of these factors undercut potential word of mouth benefits and reinscribe consumer needs for tangibility, typified by desires for a paper contract or a need to see an office.  This "seeing is believing" mentality preserves the status quo, as do gatekeepers on the fence between informality and formality.  Furthermore, South Africa is a country that is relatively heavily banked, so that mobile money is not something people need.  Moreover, there are very heavy regulations, and e-money can only be lent by banks.  In these "less than ideal payment systems," the risk of overdraft fees presents an additional deterrent to adoption.  When national policies must balance between financial inclusion and consumer protection, South Africa leans toward protection.

"Sports Betting in Uganda: Causes and Consequences" by Sylvan Herskowitz of UC Berkeley encouraged those afflicted with academic snobbery to take a "multi-billion dollar global industry" seriously, which has "exploded across sub-Saharan area" and "quintupled between 2009 and 2013," thanks to a weak regulatory environment, access to international betting markets, new technology to manage bets, and the credibility of payouts  With more than 1500 betting branches in a country with less than forty million people, Uganda is a vibrant area of economic experimentation.  He laughed at how the signifiers of betting culture were often invisible to Westerners, however, as in the case of the location of an  Ebola washing station in front of betting station in Liberia in a New York Times photograph.  One of Herskowitz's photographs documented all the international football tickets he had bought.  Like most betters he had lost his investment in all of them.  This is not surprising, since the standard multi-match ticket requires that all of the wins need to take place.  (The lure is that the winning long-shot ticket offers a large payout.

He argued that researchers need to document neglected issue, particularly one with strong behavioral biases.  For Ugandan betters this meant spending more than they normally would and ignoring how betting crowds out other expenditures.  The numbers are significant, because in his study group of betters, expenditures on betting represented a median 11% outlay of income and a mean of 15%.  Most of his respondents (75%) were heads of household.  Even though 40% of their families knew they bet, only 25% knew how much they bet.  In explaining his work on communities around Kampala on financial motivations, he dropped "economic speak" for a moment to characterize incentives as "if you want to get stuff that's big and expensive" but are constrained in ability to save or access to credit.  After all, betting is one way to generate liquidity, despite its bad rate of return.  In studying driving factors, researchers offered betters cash or betting tickets and designed the experimental situation so that timing might be before or after they got tickets. The prime increases demand for betting tickets by 15-25%.  Participants also chose higher payouts.  An experiment in which he gave them a wallet for setting money aside for betting seemed to decrease betting.  Rather than frame it as "overrationalizing an activity" he saw it as "encouraging people to reflect."  He reported a 10% reduction in betting among those who had underestimated their expenditures.  He looked forward to testing more experimental primes to sort out budgeting and failure aggregators, improving data quality by decreasing noise in the results, refining the wallet as a physical instrument, differentiating the benefits of a tangible object from simple targeting, and doing testing in relation to other expenditures such as food.

Betting Against the House: Chance, Risk and Fraud

The only all-male panel at the IMTFI focused on risky monetary behaviors.  The presenters led off with a talk about "Gambling and Mobile Money Payments: A Case Study of Sports Betting in Uganda" by Bruno Yawe and Kizito Ssengooba that focused on how value creation and value destruction might be interrelated in the field. They observed that 25% of sports betting shops were licensed (37 out of 104 enterprises) and that mobile money was used for twenty-four hour betting, which was especially popular among working class participants.  (Football was apparently the number one sport for betting activity, according to the researchers.)  

They focused their attention on six online betting shops and on questions about the regulatory environment, the addictive nature of betting, possible links to youth poverty, and mechanisms for redress in rigged and one-way gambling operations.  Some of the most dramatic narratives involved students missing graduation after tuition money had been gambled away.

"Challenges Facing the Uptake of M-Insurance Loyalty Based Life Insurance Schemes: A Case Study of the yuCover Microinsurance scheme in Kenya"  by Nelson Karani Nyachiro and Cyrus Isaboke focused on services offered by yuMobile in partnership with Jubilee Insurance and Micro Ensure,  The yuCover insurance scheme rewarded subscribers with renewable monthly life and disability coverage.  Nonetheless researchers found that belief and culture propagated biases that insurance was for old men and the owners of motor vehicle owners, and even one of the researchers confessed to being among the uninsured.

The final panel on "Automated Teller Machine Fraud in South-West Nigeria: The Shoe-Wearer's Perspective" by Oludayo Tade and Oluwatosin Adeniyi opened with a recap about the fun of participating in a conference in which economists used to crunching numbers could also hear about sorcery and comic books.  Their study focused on the life chances of victims after ATM fraud and emphasized Victims Precipitation Theory (VPT), Lifestyle Theory (LT), and Routine Activity.  The methodology mainly adopted a snowballing approach.  They discovered that the fraudsters identities tended to be kin, lovers, children, and friends, although non-kin third parties could also be victimizers.  Victims often had 2-5 ATM cards and were plagued by illiteracy, illness, or old age and weak constitutions, although even an economics professor had been victimized, according to the researchers.  Despite the protections of PIN numbers and daily limits, there was no shortage of sad story, including the tale of a man who needed help with pushing buttons who was cheated and five young people finishing national service who found their accounts emptied. Such fraud could have particularly destructive effects on students.  It could also lead to business collapse, as well as undermining the entire purpose of saving among community members.  A culture of blaming victim or of officials being unwilling to use footage from CCTV to catch perpetrators could worsen the situations.  It was also interesting to hear about the appropriations and improvisations that banks and users deployed, which included constructing screens to block views of PIN entry.

Discussant Jan Chipchase of frog design asked presenters to reflect about how their own life practices had been changed by their research.  He also pushed the question of whether "exploring topics that are taboo in shape or form or touching on taboo topics" might require researchers "to get participants' perspectives" by joining in  He even suggested that it might be "an obligation to go down that road."