Page 23 - FIGI - Use of telecommunications data for digital financial inclusion

Many of these risks are aggravated by the con- increase if alternative data becomes fundamental to
text in which telecom data is used. Subscribers credit decisions. Because alternative data is intend-
may be illiterate or unfamiliar with the services they ed to provide greater inclusion for low-income and
need. They are also likely to be using old technolo- historically disadvantaged populations, the nega-
gy and low-cost data services that cannot transmit tive consequences of systemic shocks resulting from
the quality and quantity of information consumers alternative data may disproportionately harm those
need, because most subscribers using mobile money who can least afford it.
and fintech services are not using smartphones and
the relevant technology does not have the ability to 6.4.4 Cyber security
transmit large data files such as disclosure and con- As telecommunications data becomes more sought
sent forms. For example, a study in Tanzania revealed after, cyber security will become more important,
that only 41% of customers were informed of terms both for telecommunications operators and their
and conditions when signing up for MNO-led insur- partners who are permitted to access their systems.
ance product. Standards, discussed in section 6.2, This may not be a matter only of preventing fraud
21
may be helpful in this area, failing which it may be from theft or leakage of data, but also of manipu-
necessary to introduce greater regulation to ensure lation of telecommunications data to obtain desir-
proper information is provided and consent obtained. able automated decisions made by digital financial
The weakness of consent as a justification for using service providers in reliance on such data.
personal data raises the question whether it is worth
requiring consent at all in some cases. The stronger 6.4.5 Data localisation
the social benefit of using the data, the weaker the Data localisation requirements are increasingly being
rationale is for protecting the consumer who is in legislated and enforced, often with unforeseen conse-
reality not meaningfully protected anyway. It will be quences. This and other restrictions on transporting
worth exploring whether the emphasis should rather data outside a jurisdiction may also present impedi-
be placed on building privacy protections into data ments to use of telecommunications data for greater
access in a manner that minimizes disclosure of per- financial inclusion. This is likely so particularly where
sonal data while maximizing its utility. provider groups operate in multiple jurisdictions. The
The project in Brazil described in the accompany- combined telecommunications data of a telecommu-
ing box suggests that it is indeed possible to use per- nications group active in several countries in a region
sonal data without the customer’s consent while still is likely considerably more useful than keeping the
minimizing the privacy risk to the individual. In that data in national silos. Likewise, making telecommu-
project, the credit bureaus remain key, and thus data nications data available to digital financial service
governance placed within a centralized, regulated providers and their technical partners at scale across
institutional context. Greater flexibility to make data multiple countries allows them to draw more precise
available while securing privacy might be achieved insights about their potential customers. The impor-
also through using publicly owned data collabora- tance of allowing cross-border transfer of data for
tives relying on robust frameworks setting out rights digital financial services, including for inclusion, is
and responsibilities for access to the data and the increasingly widely recognised. 22
processing that can be applied to it.
6.4.6 Identification systems
6.4.3 Financial stability and alternative data The development and adoption of standards provid-
The G20 Guidelines recognize that fintech compa- ing for use of unique identifiers across organisations
nies providing credit and insurance services have and data sets was mentioned in section 8.2. This
developed credit assessment models that use alter- might be achieved by, or coordinated with, the adop-
native data, in particular mobile phone data, to evalu- tion of national identification systems and ID numbers
ate risks. The G20 noted concerns about explicability where these are being developed. Use of unique
23
and bias, and also that these data sets have been used identifiers for telecommunications data on a national
only during the period of positive economic growth basis may also reduce risks of error to the consumer
since 2008 and therefore have not been stress-test- of digital financial services. The use of other iden-
ed against potential negative economic periods, tifiers such as MSISDN (mobile telephone numbers)
which could result in financial stability risks. However, carry high risks, particularly where numbers may be
systemic risks associated with alternative data may transferred to other people. 24

Use of telecommunications data for digital financial inclusion 21