Financial literacy or risk preference? Micro-determinants

Financial literacy or risk preference?Micro-determinants of index insurance demand

Yesuf M. Awel and Theophile T. Azomahou

Maastricht University and UNU-MERIT

y.awel@maastrichtuniversity.nl

International Conference: ”Future perspectives on Innovation and Governance inDevelopment”

26-28 November, Maastricht

November 28, 2014

Yesuf M. Awel and Theophile T. Azomahou Index insurance demand November 28, 2014 1 / 30

Outline

1 Introduction

2 Context and Data

3 Empirical Strategy

4 Empirical Results

5 Conclusion

Outline

1 Introduction

2 Context and Data

4 Empirical Results

5 Conclusion

Introduction

The literature on agricultural technology adoption is ripe inidentifying the constraints that deter adoption.

education or information, liquidity constraint, risk constraint, amongothers, (see review: Foster and Rosenzweig, 2010).

Several interventions have been in place to address the constraints.

However, interventions that could relax risk constraint has been largelymissing.

A natural response is to offer insurance.

Nevertheless, the insurance market is underdeveloped or inexistent inmost developing countries.

information asymmetries and transaction costs (Clarke and Dercon,2009; Skees, 2008).

An innovation that could address these problems is Index basedinsurance.

Weather Index Insurance

Payout in

Limit Threshold

Rainfall in

Maximum

Payout

measures a specific weather variablefor a particular product based onhistorical weather data. It thenspecifies a threshold and a limit formaking payouts.

makes payouts based on valuesobtained from an index withoutcalculating actual losses of a policyholder.

involves basis risk

Payout in

Limit Threshold

Rainfall in

Maximum

Payout

involves basis risk

Payout in

Limit Threshold

Rainfall in

Maximum

Payout

involves basis risk

Payout in

Limit Threshold

Rainfall in

Maximum

Payout

involves basis riskYesuf M. Awel and Theophile T. Azomahou Index insurance demand November 28, 2014 6 / 30

The problem: low uptake

Demand for the product is rather low.

Randomized Control Trial (RCT) based index insurance offers reporttake up rate of 20% in Ethiopia (Hill and Robles, 2010), between 6%and 36% Norton et al. (2011)

17% in Malawi (Gine and Yang, 2009),

16% in India (Cole et al., 2012).

This poses a need to understand why farmers do or do not take-upWII

Our study

Only a few studies have assessed the factors that affect demand indeveloping countries inter alia Gine et al. (2008); Cole et al. (2012);Mobarak and Rosenzweig (2012); Karlan et al. (2012)

We assess the factors that influence demand for WII.

large commercially traded index insurance in Africaelicit risk preference based on lottery choice gamemeasure financial literacy using series of questions.

Main Results: financial literacy and liquidity constraint are significantfactors that affect demand for index insurance

Outline

1 Introduction

2 Context and Data

4 Empirical Results

5 Conclusion

Context I

Formal insurance accounts for less than 1% of GDP and largely inurban Ethiopia.

The rural areas account for about 80% of the inhabitants that facemultiple perils (drought, flood, illness and so on).

Most frequent self reported shocks: drought (40.2%), illness (30.2%), death of household member (22.8%) and pest infestation(22%)(Hill et al., 2011).

Context I

Context II

The focus of this study is on the microinsurance product designed toaddress weather shock:

Piloted in 2009 in village Adiha, Tigray; later scaled to 5 more villagesin 2010.Now, covers around 79 villages (20,015 households) in Tigray and apilot roll-out village in Amhara (insuring 350 households) (OxfamAmerica, 2013)Two local insurance companiesCrops covered (Teff, Wheat, Barley and others) in two windows, earlyindex and late index (Oxfam America, 2013)The early index addresses deficit/delay of onset rainfall, the late indextargets deficit/early end of rainfallBoth windows pay once every four or five yearsFarmers have an option of purchasing insurance with cash or with labor(Insurance for Work).

Study areas

Data and Sampling

We draw stratified random sample of 169 subscribers and 106non-subscribers in 5 villages, Tigray.

The data was collected during May-June, 2013.

We filled standard household socio-economic questionnaire with focuson risk and insurance. The questionnaire includes:

modules on household demographics, household assets and wealth,detailed consumption expendituremodules on risk and time preference, financial and insurance literacy,risk and risk coping mechanismsdetails of insurance participation, insurance premium and related issues.

Eliciting preference and Measuring financial literacy I

Eliciting risk preference

Table : Lottery Choice Game

Task Safe Option (S) Risky option

pl Rl pg Rg

A 2 0.5 0 0.5 5B 2 0.5 0 0.5 10C 2 0.5 0 0.5 15

CRRA utility: U(x) = x1−r

EU(R) =

(0.5 ∗ U(Rl) + 0.5 ∗ U(Rg )

)≥ U(S) = EU(S)

Eliciting preference and Measuring financial literacy II

Measuring financial literacy:

Basic financial literacy questions adapted from Cole et al. (2011)

Insurance specific literacy questions based on Madajewicz et al. (2010).

Financial literacy questions-I

1 How much is 3+4?

2 If you have 500 Birr and friends give you 200 Birr, how many Birr do you have?

3 How much is 35+82?

4 What is 3 multiplied by 6?

5 What is one-tenth of 400?

6 Suppose you want to buy liter of oil that costs 37 birr and you have 100 Birr note. How much change will you get?

7 Suppose you borrowed Birr 1000 from a money lender and the interest rate was 2% per month. If you made norepayment for three months, how much would you owe? [1 ] less than Birr 1020 [ 2 ] exactly Birr 1020 [3 ] more thanBirr 1020 [ -999 ] DK [-9999 ] RA

8 Suppose you need to borrow Birr 1000. Two people offer you a loan. One loan requires you pay back Birr 1100 in onemonth. The second loan requires you to pay back Birr 1000 and 15 percent interest per month. Which loan would youprefer? [1 ] Birr 1100 in one month [2 ] birr 1000 and 15 percent interest [ -999 ] DK [-9999 ] RA

9 Suppose that you saved Birr 1000 in a saving account and were earning an interest rate of 4% per year. If prices wereincreasing at a rate of 2% per year, after one year, would you be able to buy more than, less than or exactly the sameamount as today with the money in the account? [1 ] less than today [2 ] exactly as much as today [ 3 ] more thantoday [ -999 ] DK [-9999 ] RA

10 Do you think the following statement is true or false? For a farmer, planting one crop is usually safer than plantingmultiple crops. [1 ] True [ 0 ] False [ -999 ] DK [-9999 ] RA

Financial literacy questions-II

1 When does insurance give you a payout?

a When your yields are poorb When rainfall is below a certain level according to rain gauge or satellitec When rainfall is below a certain level on your fieldd Other please describe [-999] DK [-9999] RA

2 Will you receive a payout every time your yields are poor?

a Yesb No [-999 ] DK [-9999 ] RA

3 If you receive an insurance payout, will the payout cover all of your losses or only a part of your losses?

a The insurance will cover all of my lossesb Most of the time the insurance will cover only part of my lossesc Other please describe [-999 ] DK [-9999 ] RA

4 Would you ever receive a refund of your premium?

a Yesb No [-999 ] DK [-9999 ] RA

5 What is the organization that is offering the insurance?

a DECSIb RESTc Oxfam Americad Nyala Insurance Co.e Africa Insurance Co.f Other please describe [-999 ] DK [-9999 ] RA

Outline

1 Introduction

2 Context and Data

4 Empirical Results

5 Conclusion

Baseline specification: Probit model

Our baseline specification is Probit model as below:

I ∗ = F (x′β + ζ) (1)

P(Ii = 1|x) = Φ(x′β + ζ > 0) (2)

where I ∗ is the latent variable of decision to buy insurance or not.

Φ is the normal cumulative density function (CDF).

x is vector of regressors that determine insurance purchase.

β is vector of parameters to be estimated using maximum likelihood.

Special Regressor Approach

To obtain consistent probit estimates in (2), all the right hand sidevariables should be exogenous.

Financial literacy literature point the possible endogeneity of thefinancial literacy variable (Van Rooij et al., 2011; Behrman et al.,2010).

Omitted variableMeasurement error.

We address the problem, using Special Regressor approach (Lewbel,2000; Lewbel et al., 2012).

Formally, the model is:

I ∗i = fliβf + xiγ + V + εi (3)

V = fliβv + xiΠ1 + ziΠ2 + Ui (4)

E (zε) = 0,E (V ) = 0,U⊥(x , fl , z , ε)

where i = 1, 2, ...,N,

fli measures financial literacy possibly endogenous variable,

xi is vector of exogenous variables,

V is the special regressor-land owned

zi is vector of additional instruments:

Frequency of extension contacts during dry seasonAccess to information (radio and mobile ownership)

Outline

1 Introduction

2 Context and Data

4 Empirical Results

5 Conclusion

Empirical results I: Descriptive statistics

Table : Descriptive statistics by insurance status

WII non-buyers WII buyers Full sampleMean Sd. Mean Sd Mean Sd

PreferenceRisk aversion parameter 0.44 0.18 0.34 0.19 0.38 0.19Time Preference (1 if discount rate ≥ 30 percent) 0.52 0.50 0.57 0.50 0.55 0.50Financial literacyBasic financial literacy index 0.44 0.24 0.51 0.20 0.48 0.22Advanced financial literacy index 0.39 0.19 0.48 0.16 0.45 0.18Demographic characteristicsAge of household head 47.0 14.1 43.3 12.1 44.8 13.0Household Size 5.34 2.30 5.26 1.88 5.29 2.05Sex of household head 0.37 0.48 0.40 0.49 0.39 0.49Religiosity (1=attend daily) 0.22 0.42 0.15 0.36 0.18 0.38Head Illiterate (1=yes) 0.76 0.43 0.58 0.50 0.65 0.48Head some elementary education (1=yes) 0.21 0.41 0.33 0.47 0.28 0.45Head some secondary education (1=yes) 0.029 0.17 0.093 0.29 0.068 0.25Other Risk management strategiesIddir Participation (1=yes) 0.11 0.31 0.16 0.37 0.14 0.35Eqqub Participation (1=yes) 0.038 0.19 0.11 0.32 0.083 0.28PSNP Participation (1=yes) 0.52 0.50 0.69 0.46 0.62 0.49Remittance & other gifts received 0.17 0.38 0.093 0.29 0.12 0.33Credit (1=accessed) 0.43 0.50 0.42 0.50 0.43 0.50Has formal Saving (1=yes) 0.067 0.25 0.13 0.34 0.11 0.31Location dummiesAdiha (1=yes) 0.11 0.31 0.28 0.45 0.21 0.41Awetbikalsi (1=yes) 0.17 0.38 0.18 0.39 0.18 0.38Genetie (1=yes) 0.29 0.46 0.099 0.30 0.17 0.38Hadealga (1=yes) 0.27 0.45 0.14 0.35 0.19 0.39Hadushadi (1=yes) 0.16 0.37 0.30 0.46 0.25 0.43

Empirical results II: SR, CF and Probit estimates

Table : Marginal Effects: Special regressor, Control function and probit estimates

Special regressor Control function Probit

Land owned 0.089∗∗∗ 0.007 0.006(0.028) (0.013) (0.014)

Basic financial literacy index 0.784∗∗∗ 0.871∗ -0.109(0.219) (0.490) (0.150)

Risk aversion parameter -0.146 -0.107 -0.208(0.171) (0.351) (0.376)

Time Preference (1 if discount rate >= 30 percent) 0.124∗∗∗ 0.030 0.048(0.046) (0.054) (0.057)

PSNP Participation (1=yes) 0.129∗ 0.138∗∗ 0.111∗

(0.072) (0.057) (0.061)

Eqqub Participation (1=yes) 0.102∗∗∗ 0.273∗∗ 0.314∗∗∗

(0.033) (0.118) (0.114)

Iddir Participation (1=yes) 0.018 0.134∗ 0.137∗

(0.046) (0.081) (0.079)

Shock dummy (1= if household experienced shock last 3 years) 0.085∗∗∗ 0.048 0.003(0.032) (0.112) (0.121)

Other controls Yes Yes YesLocation dummies Yes Yes Yes

Observations 243 268 274

Standard errors in parentheses∗ p < 0.1, ∗∗ p < 0.05, ∗∗∗ p < 0.01

Land owned 0.089∗∗∗ 0.007 0.006(0.028) (0.013) (0.014)

(0.072) (0.057) (0.061)

(0.033) (0.118) (0.114)

(0.046) (0.081) (0.079)

Land owned 0.089∗∗∗ 0.007 0.006(0.028) (0.013) (0.014)

(0.072) (0.057) (0.061)

(0.033) (0.118) (0.114)

(0.046) (0.081) (0.079)

Empirical results III: SR with alternative financial literacyand risk preference measures

Table : Special Regressor Estimates

(1) (2) (3) (4) (5)

Basic financial literacy index 0.432∗

(0.248)

Standardized basic financial literacy index 0.173∗∗∗

(0.048)

Advanced financial literacy index 0.822∗∗

(0.404)

Basic financial literacy PRIDIT score 1.789∗∗∗

(0.678)

Advanced financial literacy PRIDIT score 1.575(1.023)

Risk aversion parameter -0.146 -0.130 -0.104 -0.071(0.171) (0.236) (0.161) (0.148)

log risk attitude based on hypothetical inv’t question 0.005(0.003)

Land owned 0.089∗∗∗ 0.058∗∗ 0.055∗ 0.040 0.060∗∗

(0.028) (0.028) (0.029) (0.033) (0.030)Other regressors Yes Yes Yes Yes YesLocation dummies Yes Yes Yes Yes Yes

Observations 243 243 243 243 241

(1) (2) (3) (4) (5)

(0.248)

(0.048)

(0.404)

(0.678)

Land owned 0.089∗∗∗ 0.058∗∗ 0.055∗ 0.040 0.060∗∗

Observations 243 243 243 243 241

(1) (2) (3) (4) (5)

(0.248)

(0.048)

(0.404)

(0.678)

Land owned 0.089∗∗∗ 0.058∗∗ 0.055∗ 0.040 0.060∗∗

Observations 243 243 243 243 241

Outline

1 Introduction

2 Context and Data

4 Empirical Results

5 Conclusion

Key results and Implications

Risk preference:

efforts should be made to reduce possible basis risk of the product.

Financial literacy:

strengthening efforts in provision of financial literacy programs.

Liquidity constraint:

better designs that account for liquidity constraint of the households.sales of the insurance policy could be done immediately after theharvest season.effecting payouts as immediately as possible once the index triggers thepayment.

Thank You!

Bibliography I

Barrios, S., Bertinelli, L., and Strobl, E. (2010). Trends in rainfall and economic growth in africa: A neglected cause of theafrican growth tragedy. The Review of Economics and Statistics, 92(2):350–366.

Behrman, J. R., Mitchell, O. S., Soo, C., and Bravo, D. (2010). Financial literacy, schooling, and wealth accumulation.Technical report, National Bureau of Economic Research.

Clarke, D. and Dercon, S. (2009). Insurance, credit and safety nets for the poor in a world of risk. United Nations, Departmentof Economics and Social Affairs, Working Paper, 81.

Cole, S., Gine, X., Tobacman, J., Townsend, R., Topalova, P., and Vickery, J. (2012). Barriers to household risk management:evidence from india. Harvard Business School Finance Working Paper, (09-116):104–35.

Cole, S., Sampson, T., and Zia, B. (2011). Prices or knowledge? what drives demand for financial services in emerging markets?The Journal of Finance, 66(6):1933–1967.

Dell, M., Jones, B. F., and Olken, B. A. (2008). Climate shocks and economic growth: Evidence from the last half century.NBER Working Paper, 14132.

Dercon, S. (2004). Growth and shocks: evidence from rural ethiopia. Journal of Development Economics, 74(2):309–329.

Gine, X., Townsend, R., and Vickery, J. (2008). Patterns of rainfall insurance participation in rural india. The World BankEconomic Review, 22(3):539–566.

Gine, X. and Yang, D. (2009). Insurance, credit, and technology adoption: Field experimental evidencefrom malawi. Journal ofDevelopment Economics, 89(1):1–11.

Hill, R., Hoddinott, J., and Kumar, N. (2011). Adoption of weather-index insurance: Learning from willingness to pay among apanel of households in rural ethiopia. International Food Policy Research Institute Discussion Paper, (1088).

Hill, R. and Robles, M. (2010). Flexible insurance for heterogeneous farmers: results from a small scale pilot in ethiopia.International Food Policy Research Institute Working Paper.

Karlan, D., Osei, R. D., Osei-Akoto, I., and Udry, C. (2012). Agricultural decisions after relaxing credit and risk constraints.Technical report, National Bureau of Economic Research.

Bibliography II

Lewbel, A. (2000). Semiparametric qualitative response model estimation with unknown heteroscedasticity or instrumentalvariables. Journal of Econometrics, 97(1):145–177.

Lewbel, A., Dong, Y., and Yang, T. T. (2012). Comparing features of convenient estimators for binary choice models withendogenous regressors. Canadian Journal of Economics/Revue canadienne d’economique, 45(3):809–829.

Madajewicz, M., Peterson, N., Gebregziabher, G., and Awel, Y. M. (2010). How effective is index insurance in helping farmersto manage climate risk, build resilience, and improve their livelihoods? Technical report, Project funded by Oxfam Americaand led by the International Research Institute for Climate and Society, Columbia University.

Mobarak, A. and Rosenzweig, M. (2012). Selling formal insurance to the informally insured. Yale University Economic GrowthCenter Discussion Paper, (1007).

Norton, M., Holthaus, E., Madajewicz, M., Osgood, D. E., Peterson, N., Gebremichael, M., Mullally, C., and Teh, T. (2011).Investigating demand for weather index insurance: Experimental evidence from ethiopia. In Annual Meeting Agricultural andApplied Economics Association, Pennsylvania.

Oxfam America (2013). R4 Rural Resilience Inititative Quarterly Report: January-March 2013. Technical report.

Skees, J. R. (2008). Innovations in index insurance for the poor in lower income countries. Agricultural and Resource EconomicsReview, 37(1):1.

Van Rooij, M., Lusardi, A., and Alessie, R. (2011). Financial literacy and stock market participation. Journal of FinancialEconomics, 101(2):449–472.

Financial literacy or risk preference? Micro-determinants

Documents

DETERMINANTS DETERMINANTS OF VENTRICULAR FUNCTION DETERMINANTS DETERMINANTS OF VENTRICULAR FUNCTION STROKE VOLUME PRELOAD CONTRACTILITY CARDIAC OUTPUT

Increasing Statistical Literacy among Students through the Census at School: Food Preference Survey Presented by Suzanne King, Public Affairs Office Director

Adult Literacy and Postsecondary Education Students ... · Literacy skills are also important determinants of individuals' labor market outcomes. At given levels of education, increasing

Determinants of health information literacy in vocational ... · • Prevention of and successful psychosocial adjustment to illness (Jung, 2014). Especially in emerging adulthood:

RANCANG BANGUN APLIKASI FINDING-TUTOR BERBASIS …repository.its.ac.id › 43101 › 1 › 5113100079-Undergraduate_Thesis.pdfgender preference, age preference, lesson preference,

Economic evaluation, value of life, stated preference ...372086/FULLTEXT01.pdfstated preference methodology and determinants of risks. Publisher: Örebro University 2010 trycksaker@oru.se

Attitudinal Determinants of Son Preference for Pakistani ...which results in high mortality among female infants in Pakistan (Arnold, 1992). It isfactor that shows bias towards female

67.Preference-BasedMultiobjectiveParticleSwarm Preference-B

A Preference Reference: Common Issues that Arise in Delaware Preference Litigation€¦ · · 2017-04-15A Preference Reference: Common Issues that Arise in Delaware Preference Litigation

Determinants of Financial Literacy and Use of Financial ...€¦ · ‘unorganized worker’ have been clearly defined in the Unorganised Workers’ Social Security Act (2008) which

Determinants of Music Preference (Bestimmungsgrößen für

Mental health literacy measures evaluating knowledge, attitudes … · 2017. 8. 25. · barrassment, problems in symptom identification and a preference for self-reliance were the

The economic determinants of US migration policy · 4 at all towards district level economic determinants. Goldin’s (1994) study of the introduction of the literacy test provision

Increasing Statistical Literacy among Students through the Census at School: Food Preference Survey

Food Truck Parking Location Assignments - George Mason · 2015. 5. 12. · Preference Totals Location 1st Preference 2nd Preference 3rd Preference Total Farragut Square 17th St 195

Tariff Preference and Trade Cost as Determinants of Export ...Tariff Preference and Trade Cost as Determinants of Export: The Case of Bangladesh ... The first part of the study aims

Addressing the Social Determinants of Health: The Role of Health Literacy - Hawaiilabor.hawaii.gov/ola/files/2013/08/Language-Access... · 2013. 8. 19. · Addressing the Social Determinants

A coenzyme - PNAS · mutant enzyme with kinetic characteristics and coenzyme preference comparable to naturally occurring homologous NAD-dependentenzymes. Descriptions ofthe determinants

Preference Uncertainty, Preference Learning and Paired

The Determinants of Literacy and Numeracy, and the Effect ... · In section 2 the determinants of literacy and numeracy are considered. A discussion of the data base and variables