Page 151 - Special Topic Session (STS) - Volume 1

STS422 Damola Owalade
5
The data quality assessment from the two case studies based on the
methodology has discussed in section 2 are presented below.
6
Zimbabwe
7
The data was on loans from microfinance entities, department stores,
other retailers and agricultural suppliers.
o Relevance
The data is relevant as a sample source, but it is limited to one of the four
credit bureaus in Zimbabwe including the Reserve Bank of Zimbabwe’s Credit
Reference Bureau (CRB). Notably, the credit records do not include retail credit
from commercial banks and mobile money operators. There is a dearth of
publicly available data on the credit market in Zimbabwe but according to
Zimbabwe FinScope Consumer Survey 2014, commercial bank credit was
accessed by only 4% of the adult population compared to the non-bank credit
which was accessed by 10% of the adult population. If it is assumed that the
ratio of bank to non-bank credit, based on Zimbabwe FinScope 2014, still
holds, then the relevance of the data will be limited to only a segment of the
formal credit market. However, the data is adequate to inform policy design
for credit users from microfinance banks and institutions.
o Accessibility
The data was provided to the i2i team in a spreadsheet. There was no
engagement with the management information system of the credit bureau.
A list of selected potential survey respondents were provided with names and
contact details. Administrative data was analysed at the office of the credit
bureau in Harare in adherence to data privacy and sharing protocols.
o Interpretability
The research team depended on the contact person from the credit bureau
to explain the data as there was no data dictionary. In ensuring high quality
data, it is required that interpretation of data is informed by institutional based
definitions of the data points as outlined in a data dictionary.
o Coherence
The data is stable and there were no changes to the data template over the
period of study. Given data privacy concerns and time availability, the data
could not be matched with actual records from the credit providers.

5 Data Quality Assessment Tool for Administrative Data, Iwig W, Berning M, Marck P, Prell M,
February 2013 available at https://www.bls.gov/osmr/datatool.pdf
6 The research project was conducted between July and October 2017. Fieldwork was
conducted by Research Continental-Fonkon (RCF)
7 The indicators provided by the credit bureau include Credit application, Accepted and
rejected applications (reasons for rejection), Loan value, Credit provider type (Microfinance
banks and institutions, retail stores, and agriculture input suppliers), Use case of credit (e.g.
consumption or productive categories) and Repayment defaults

140 | I S I W S C 2 0 1 9