Saving More in Groups: Field Experimental Evidence from Chile

Out­line

Peers can have both neg­a­tive and pos­i­tive ef­fects on hu­man be­hav­ior. The use of peers as a com­mit­ment de­vice to reach a shared but in­di­vid­ual goal is a wide­ly ob­served phe­nom­e­non, both in­for­mal­ly (e.g., run­ning groups or study groups) and for­mal­ly (e.g., weight-loss groups). On the oth­er hand, peer pres­sure can also lead in­di­vid­u­als to make sub-op­ti­mal de­ci­sions. While self-help peer groups have been sub­ject to the­o­ret­i­cal analy­sis, to our knowl­edge there is no clean ev­i­dence in­ves­ti­gat­ing whether peer groups make par­tic­i­pants more like­ly to in­crease their sav­ings, and if so, what as­pects lead to this ef­fect.

Self-help peer groups in­crease sav­ings

For this project, we con­duct­ed two ran­dom­ized field ex­per­i­ments among low-in­come mi­cro-en­tre­pre­neurs in Chile to study the pow­er of 1) self-help peer groups and 2) peer-re­lat­ed text mes­sages for pre­cau­tion­ary sav­ings. We of­fered 2,687 mi­cro-en­tre­pre­neurs, who met reg­u­lar­ly in groups as clients of a mi­cro­cre­d­it as­so­ci­a­tion, the op­por­tu­ni­ty to open a for­mal sav­ings ac­count. Our first ex­per­i­ment, the “Peer Group Ex­per­i­ment,” shows that self-help peer groups, which com­bine pub­lic goal set­ting, mon­i­tor­ing in the peer group and recog­ni­tion for those who com­ply, sig­nif­i­cant­ly in­crease sav­ings. A sec­ond ex­per­i­ment, con­duct­ed a year lat­er, tests an al­ter­na­tive de­liv­ery mech­a­nism and shows that ef­fects of sim­i­lar size can be achieved through feed­back text mes­sages, with­out meet­ings, re­wards, or peer pres­sure.

Our re­sults ad­dress a larg­er point about be­hav­ioral in­ter­ven­tions ver­sus fi­nan­cial in­cen­tives to af­fect be­hav­ior. A grow­ing pat­tern of ev­i­dence shows the im­por­tance of so­cial in­cen­tives and sur­pris­ing­ly lim­it­ed ef­fects of mon­e­tary re­wards, even for fi­nan­cial de­ci­sions. Our find­ing of strong ef­fects in the Peer Group Treat­ment and sug­ges­tive ev­i­dence from the feed­back mes­sages, com­pared to the lim­it­ed ef­fec­tive­ness of the in­ter­est rate, fits into this pat­tern. While tra­di­tion­al eco­nom­ic in­cen­tives may be ef­fec­tive in con­texts where in­di­vid­u­als lack mo­ti­va­tion, they may have lim­it­ed im­pact if the con­straint that im­pedes the be­hav­ior change lies else­where. Im­ple­ment­ing be­hav­ior change can be chal­leng­ing, even for mo­ti­vat­ed in­di­vid­u­als – ei­ther psy­cho­log­i­cal­ly, due for ex­am­ple to self-con­trol prob­lems, or prac­ti­cal­ly, due for ex­am­ple to com­pli­cat­ed process­es. In these sit­u­a­tions, poli­cies that fa­cil­i­tate com­pli­ance may be more ef­fec­tive than poli­cies that fur­ther in­crease in­cen­tives.

Re­search Team

Author

Dina Pomeranz

Assistant Professor of Microeconomics, endowed by the UBS Center

Zurich ZCED

Fe­lipe Kast

Pon­ti­f­i­cia Uni­ver­si­dad Católi­ca de Chile

Stephan Meier

Co­lum­bia Uni­ver­si­ty GSB and NBER