Assessing the Quality of Self-Reported Financial Information

Carlos Madeira, Paula Margaretic, Felipe Martínez, Pedro Roje

Research output: Contribution to journalArticlepeer-review

1 Scopus citations


It is well known that self-reported financial information differs from administrative records. This article advances previous studies by using a unique matched representative sample of individual borrowers from the Chilean Household Finance Survey with administrative banking loan records. Our linked dataset allows us to test whether the differences between the two sources are due to the number of nonreported loans or to differences in the reported loan amounts. We show that discrepancies in debt ownership are larger when respondents have a mortgage and are not the highest income member of the family, when respondents have weak financial literacy and do not have confidence in the study, and when respondents live in complex or multigenerational households. However, borrowers report the maturity of their mortgage and installment loans quite accurately. Concerning loan amounts, differences between the two sources decrease with the financial literacy of the respondent. Finally, a simple form of rounding can go a long way towards explaining differences in the middle part of the debt amount and maturity distributions.

Original languageEnglish
Pages (from-to)1183-1210
Number of pages28
JournalJournal of Survey Statistics and Methodology
Issue number5
StatePublished - 1 Nov 2022
Externally publishedYes


  • Consumer credit
  • Default
  • Household finance surveys
  • Measurement error
  • Mortgages


Dive into the research topics of 'Assessing the Quality of Self-Reported Financial Information'. Together they form a unique fingerprint.

Cite this