What Determines Consumer Complaining Behavior? - Davesh ...

Introduction

What Determines Consumer Complaining Behavior?

Devesh Raval

Federal Trade Commission

December 2016

Raval (FTC) What Determines Consumer Complaints? December 2016 1 / 29

Introduction

Disclaimer: The opinions expressed here are those of the authors and not necessarily those of the Federal Trade Commission or any of its Commissioners.


Introduction


Who complains to the Consumer Sentinel?

What do they complain about?

How do complaint rates compare to victimization rates?


Introduction






Introduction

Consumer Demographics

Consumer Zip Code matched to ACS 2008-2012 Demographics � Race: Percent Black, Percent Hispanic, Percent Asian � Culture: Percent College Graduates, Degree of Urbanization � Cost of Time: Median Household Income, Unemployment Rate,

Median Age, Household Size

Examine how per capita complaint rate for Consumer Sentinel varies with demographics

Data from 2015, excluding identity theft


Cl)

a. ~4 a. 0 0 o_ ... ..... Cl)

a. 3

~ ell c: '(ij

a. E 0 (.) 2

0 25 50 Population Share

Variable - Black - Hispanic

75 100

Introduction

Do minority areas complain at different rates?


5.0

Cl)

a. 0 Cl) a. 845 c_ ... .... Cl)

a. J!l c ell c '(ij 4.0 a. E 0 (,)

3.5 0 25 50

Population Share

Variable - Percent Urban

75 100

Introduction

Do urban areas complain at different rates?


Cl)

"ii 12 0 Cl) a. 0 0 o_ ... .....

9 Cl)

a. J!l c ell c '(ij

"ii 6 E 0 (,)


Variable - Percent College Educated

75 100

Introduction

Do college educated areas complain at different rates?


Introduction

How do complaint rates vary across areas?

Higher rates for black, college educated, higher unemployment areas

Lower rates for Hispanic, rural, greater HH Size areas

Small differences for median age and income


Introduction






0.20

J!l c: "iij

~015 0

(.) -0

~ Ill i7i 0.10

0.05


75

category - Auto - Bank - DebtCollection - Imposter - Prize

100

Introduction

Minority areas complain about different issues (% Black)


'i' 2.0 .!:! Cij

E 0 ~

~ 15 ·; ii E 0 u -0

~10 .c II)


75


100

Introduction

Minority areas complain about different issues (% Black)


'ti' .~ Cij

E

2.0

0 ~1.5

.l!l c: ·; ii E 0 u 010 ~ ca .c: II)


75


100

Introduction

Minority areas complain about different issues (% Hispanic)


'ti' .~ Cij

E

2.0

0 ~1.5

.l!l c: ·; ii E 810 -0

~ ca .c: II)

0.5


75


100

Introduction

Urban areas complain about different issues



75


100

Introduction

College educated areas complain about different issues


Introduction






�

�

Introduction

Why do consumers complain?

Higher rates of consumer complaints could reflect: Higher Propensity to Complain

Worse Consumer Experience

Typically difficult to disentangle these two stories


Introduction

I separate the two using a set of legal cases

Victim Datasets matched to Consumer Sentinel Network complaints

Linked to demographics at zip code level

Do policymakers learn about problems affecting all communities?


Introduction

For four cases, I can compare victims to complaints

Case Number of Victims Number of Complaints Case B 12,311,307 4,271 Ideal Financial 2,010,169 1,403 Platinum Trust 69,576 510 WinFixer 304,493 1,062


Analysis

Complaint rates are correlated with victim rates

Case Effect Case B 0.125

(0.033) Ideal Financial 0.151

(0.026) Platinum Trust 0.155

(0.023) WinFixer 0.173

(0.036)


~ Q) N 'fij6 E 0 ~ c: 0 ·-; g_4 0 a..

0 0 o_

Qi ~ E ·.;::; 0

> 0 25 50

Population Share 75

Case - Case B - Ideal - Platinum - WinFixer

100

Analysis

Victim rates vary widely by case (% Black) ...


~ .~1.50 iii E 0 z -;125 0

~ ::I a. ~1.00 0 0 o_

&15 UJ E ·.;::; 0

> 0.50 --.---~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~----.--'


75


100

Analysis

Victim rates vary widely by case (% Hispanic) ...


~1.6 "Cl Q)

.!::! iii E 0 z ';14 0

~ ::I a. 0 a..

0

8_12

Qi a. UJ E ·.;::; 0

>10


75


100

Analysis

Victim rates vary widely by case (% Urban) ...


Qi a. UI0.5 E ·.;::; 0

> 0.0 ->--~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-.--'


75


100

Analysis

Victim rates vary widely by case (% College) ...


S' Q)

.!::! ca E ~12 VI E ()

> g o.a o_

Qi 0.. VI E ·m0.4 0.. E 0

CJ


75


100

Analysis

... but complaint rates lower in minority areas (% Black)


25 50 Population Share

75


100

Analysis

... but complaint rates lower in minority areas (% Hispanic)



75


100

Analysis

... but complaint rates higher in urban areas


S' Q)

.!::! ca E7.5 0 6 VI E .°25.0 > 0 0 o_

&.5 VI E "iii 0.. E 800


75


100

Analysis

... but complaint rates higher in college areas


�

�

�

�

�

Conclusion

How can we account for differences in complaint propensity?

Policy answer - Outreach Contact groups that typically complain less FTC: Outreach events Marketer: Surveys / Incentives?

Statistical answer - Weighting Overweight complaints from groups that complain less But: Need data on consumer experience to construct weights


What Determines Consumer Complaining Behavior? - Davesh ...

Documents