Page 427 - Special Topic Session (STS) - Volume 3

STS552 João Falcão Silva et al.
analysed within one country, or across/between countries aiming to compare
the same statistical data under a dual perspective (eg creditors versus
debtors). Falcão Silva, João & Pradhan, Swapan-Kumar (2018) demonstrated
the importance of mirror data to enhance statistical quality as well as coverage
of data across comparable statistical domains.
In the absence of data confidentiality restrictions, mirror data exercise is
an important tool to improve the quality of the data, fill-in data gaps and
reduce bilateral asymmetries. In the case of households, mirror data exercises
can perform better estimates of their financial assets/liabilities because
households do not disclose amount/location of their cross-border positions
4
(assets/liabilities) either directly or through a survey . As confidentiality
constraints of bilateral data at granular level prevents knowing banks’ location
(BIS reporting countries), we also provide an estimation method at an
aggregate level. Furthermore, we demonstrate that our methods provide
better estimates of the households’ cross-border positions in a given country,
when bilateral data are disclosed for majority of reporting countries. We apply
our methodology to the Portuguese data as a country-case example.
Finally, as the financial assets/liabilities of households is considered
commonly as a statistical gap in the balance of payments/international
investment position (BoP/IIP) and the rest of the world (RoW) accounts
compilation, this approach could also support these two statistical domains.

2. Compilation of data for households sector – main challenges
The compilation of data for the household sector raises some difficulties
related to the data availability and accuracy. One of the main issues is the
accountability because there is no full set of accounts or ability to draw up sets
of accounts for household sector . Households’ surveys constitute one source
5
to surpass this issue. Nevertheless, some drawbacks are associated with non-
responses, estimation or underreporting of their financial assets and income.
System of National Accounts (SNA) 2008 states an example associated with
6
people earning income arising from illegal activities who may be very reluctant
to provide this information and may choose not to participate in the survey.
Similarly, it is common for households at the very top/bottom of the
distribution to be omitted from the survey either by design or on the grounds
of practicality. Low frequency and long lag in data availability are other critical
issues. Currently, there exists a lack of regular information on the households’
assets and liabilities broken down by financial instrument types such as

4 Such information from creditor/debtor sources (banks where their deposits are located)
would be complementary source for the purpose
5 that includes also non-profit institutions serving households.
6 paragraph 24.24.
416 | I S I W S C 2 0 1 9