Consump(on Smoothing, Frequency of Benefit Payments, and Effec(veness of Social Programs

Transcription

1 Consump(on Smoothing, Frequency of Benefit Payments, and Effec(veness of Social Programs Emma Aguila (USC & RAND) Arie Kapteyn (USC) Francisco Pérez- Arce (RAND) June 26, 2015

2 Frequency of Payment may MaMer for a Variety of Reasons Individuals o4en don t smooth consump=on over a pay- cycle. (eg. Mastrobuoni and Weinberg, 2007) à Food (and other types of ) consump=on may suffer at the end of the pay- cycle Less frequent but larger disbursements may facilitate investments But may cause problems with holding large amounts of cash (eg. Dupas and Robinson, 2011 )

3 Our Approach We take advantage of a Randomized Controlled Trial of a non- contributory pension program with high frequency (monthly) disbursements & the introduc=on of a similar, but lower frequency (bi- monthly) program We show that expenditures depend partly on =me elapsed since last disbursement We show there are different effects of each of these programs depending on the type of expenditure

4 Outline Background Empirical Strategy Results

5 Mexico Does Not Provide Universal Social Security Benefits Public sector employees Private sector employees Self- employed who choose to contribute Formal employment sector 42% Receive social security and health care benefits

6 Mexico Does Not Provide Universal Social Security Benefits Public sector employees Private sector employees Self- employed who choose to contribute Formal employment sector 42% Informal employment sector 58% Workers at non- compliant small firms Self- employed workers who do not contribute Receive social security and health care benefits Some receive non- contributory pension benefits and health care provided by government

7 This Study Uses the Introduc(on of Two Non- Contributory Pension Programs Reconocer: A program by the State of Yucatan 70 y Más: A Federal Program financed by the Ministry of Social Development of the Federal Government.

8 Reconocer is the first Randomized Controlled Trial of Cash Transfers for Elderly Benefits for adults age 70+ $550 pesos per month ($78 USD PPP) Treatment and control groups: experimental design Started in 2008 and ongoing Ac=ve partnership with the Government of Yucatan Funded by Na=onal Ins=tute on Aging (NIA) RAND Government of the State of Yucatan

9 Reconocer Program Was Phased into Three Yucatan Towns with Popula(ons of 20,000+ Yucatan Experiment 1 (2008) Treatment: Valladolid Control: Motul Mexico Motul Valladolid

10 Two Towns in Yucatan were Selected with Similar Characteris(cs One (Valladolid) was chosen to be treatment town and the other (Motul) was selected as control. Baseline data collec=on in August/September 2008 in Valladolid; November 2008 in Motul. Pensions in Valladolid started in December First follow- up of data collec=on in July/September in both towns.

11 70 y Más is a Federal Program that originally benefited only small towns 70 y Mas is also a non- contributory pension program Disburses Mx$1,000 pesos every two months 70 y Más originally covered towns with less than 20 thousand inhabitants 70 y Más expanded to towns with more than 20,000 (but less than 30,000) in the summer of > benefited Motul but not Valladolid. Currently the program is being further expanded to urban areas

12 Outline Background Empirical Strategy Results

13 Motul Started Receiving 70 y Mas, Roughly six Months a`er the Treatment Started in Valladolid Main Difference was the Frequency of Payment: Treatment Program Provides Mx$550 monthly versus Mx$1000 bimonthly in the control town à Goal of study is to es=mate the impact of each program à Difference amributed to Frequency of Payment

14 Timeline and Empirical Strategy Treatment (monthly) Program Starts Control (bi- monthly) Program Starts M = 0 M = 6 M = 18 W1 W2 W3 1. (W3 W1)Monthly (W3 W1)Bi- Monthly Differen(al Impact of Monthly P 2. (W2 W1)Monthly (W2 W1) Bi- Monthly Impact of Monthly Program 3. - [W3 W2]Monthly (W3 W2)Bi- Monthly] Impact of Bi- Monthly Program

15 Outcome Variables Basic needs: Health care/food consump(on Health Expenditures: Visited a Doctor, number of doctor visits, had a serious health problem but did not visit a doctor because of money, did not buy medicines because they were too expensive Expenditure on Durable and luxury goods Equipment and investment goods: Owning refrigerator, telephone, cell phone, bicycle; owns no piece of equipment; how many chickens, pigs, turkeys? Private transfers: family and other social network transfers In- Transfers to the elderly: Received money, total money received, family or friends pay for expenses; received in kind, total in kind received; received free food, food from charity Sa(sfac(on Income: Sa=sfied with family income Food (basic cons.): Luxury goods: How o4en Some=mes does does one eat food items (non- not have enough to eat, o4en cereal): meat, fruit, eggs, runs out of food, respondent diary; how much spent or someone in hhd skipped or on beverages, how much cut meals, o4en eats less than Spent on food and beverages should, O4en hungry because away from home cannot afford food, Out- Transfer from the elderly to the rest of the family: Give money, total money given, give in kind, total given in kind Family and Social Rela(onships: Sa=sfied with rela=on family members, sa=sfied with social contacts, have you been abused

16 Outline Background Empirical Strategy Results

17 Expenditures on Food at Home Monthly Program expenditure during the last 7 days days since last disbursement Bimonthly Program days since last disbursement

18 Total Expenditures Monthly Program expenditure during the last 7 days days since last disbursement Bimonthly Program days since last disbursement

19 Selected Outcome Variables

20 Doctor s Visits 0.6 Visited a Doctor Last Month Valladolid Motul 0.2 Difference W1 W2 W3-0.1

21 Work 0.2 Work for Pay Valladolid 0.05 Motul Difference 0 W1 W2 W

22 Cell Phones 0.3 Owning a Cell Phone Valladolid Motul Difference 0 W1 W2 W

23 Bicycles 0.4 Owning a Bicycle Valladolid 0 W1 W2 W3 Motul Difference

24 Transfers from Family Members 300 Total money received (MXN$) Valladolid Motul Difference 0 W1 W2 W

25 Transfers to Family Members 25 Total money given (MXN$) W1 W2 W3 Valladolid Motul Difference

26 Abuse How o`en do you feel abused? never- always (1-4) Valladolid Motul 0.4 Difference W1 W2 W3-0.2

27 Conclusions Monthly benefit is more effec=ve in improving food availability and decreasing hunger Bi- monthly benefit increased consump=on of a couple of durable goods (cell phones and bicycles) Monthly benefit decreases cash transfers from family members Bi- monthly benefit increases transfers from the older person to family members somewhat Bi- Monthly Benefit increases fear of losing money and possibly increases tensions with family members

28 Frequency MaMers!