Page 303 - Special Topic Session (STS) - Volume 1

STS441 David B. et al.

Completing the securities picture: Integrating
official securities Statistics with regulatory
1
trading data
David Buckmann, Tobias Cagala, Alena Wabitsch
Deutsche Bundesbank, Frankfurt, Germany

Abstract
With an ongoing integration of financial markets, the interconnectedness of
market participants and corresponding risks to financial stability become
increasingly important for policymakers and regulators. To develop a more
comprehensive picture on exposures to risks and relationships between
market participants, we integrate two data sources: Official statistics on
securities holdings of banks in Germany (SHS) and regulatory
transaction-by-transaction trading data that comprises millions of securities
transactions per day (MiFID). Because the datasets provide information on
stock markets from different perspectives – the SHS data show securities
portfolios whereas the MiFID data show securities transactions – integrating
the datasets is a challenge. To overcome this challenge, we combine
supervised and unsupervised machine learning algorithms and develop a
simple and transparent set of rules for the integration of the datasets. We find
that, in combination with expert heuristics, our data driven approach allows
for a successful isolation of subsamples that can be matched accurately
between the datasets. For these subsamples, the integration of both data
sources allows us analyse the exposure of banks to portfolio risks at any point
in time. This is a considerable advancement from the monthly information on
portfolios in SHS and the lack of information on portfolios in MiFID data.

Keywords
Securities Transactions; Securities Holdings; Banking Networks; Data
Integration

JEL: L14, C8

1. Introduction
Financial markets play a central role in the efficient allocation of scarce
resources. The downside of their pivotal role is that failures of financial markets
pose a threat to economic stability. To safeguard financial stability and to
improve our understanding of the ways in which financial markets work,
central banks exploit statistical and regulatory data sources. Often, these data

1 The paper represents the authors’ personal opinions and does not necessarily reflect the
views of the Deutsche Bundesbank or its staff.
292 | I S I W S C 2 0 1 9