Copyright © 2009 John Wiley & Sons, Ltd. Intell. Sys. Acc. Fin. Mgmt. 16, 5–19 (2009)
DOI: 10.1002/isaf
AUTOMATED EXPLANATION OF FINANCIAL DATA 5
INTELLIGENT SYSTEMS IN ACCOUNTING, FINANCE AND MANAGEMENT
Intell. Sys. Acc. Fin. Mgmt. 16, 5–19 (2009)
Published online 30 March 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/isaf.290
Copyright © 2009 John Wiley & Sons, Ltd.
AUTOMATED EXPLANATION OF FINANCIAL DATA
H. A. M. DANIELS
a,b
AND E. A. M. CARON
b
a
Center for Economic Research, Tilburg University, PO Box 90153, 5000 LE, Tilburg, The Netherlands
b
Erasmus Research Institute of Management, Erasmus University Rotterdam, PO Box 1738, 3000 DR, Rotterdam,
The Netherlands
SUMMARY
We describe a methodology for explanation generation in financial knowledge-based systems. This offers the
possibility to generate explanations and diagnostics automatically to support business decision tasks. The central
goal is the identification of specific knowledge structures and reasoning methods required to construct com-
puterized explanations from financial data and models. A multistep look-ahead algorithm is proposed that deals
with so-called cancelling-out effects, which are a common phenomenon in financial data sets. Our method is
an extension of the traditional variance decomposition in accounting. The method was tested on a case-study
conducted for Statistics Netherlands involving the comparison of financial figures of firms in the Dutch retail
branch. Copyright © 2009 John Wiley & Sons, Inc.
* Correspondence to: H. A. M. Daniels, Center for Economic Research, Tilburg University, PO Box 90153, 5000 LE, Tilburg,
The Netherlands. E-mail: daniels@uvt.nl
1. INTRODUCTION
Competition benchmarking or interfirm comparison (IFC) is defined as the regular measuring and
comparing of a company’s performance against its competitors or historic averages. By comparing
the financial variables of a company with those of other companies, the company can assess its
performance against objective standards and see where the company is strong or weak. Currently,
the diagnostic process for IFC is mostly carried out manually by bankers, accountants and business
consultants. The analyst has to explore large data sets in the domain of business and finance to spot
firms that expose exceptional behaviour compared with some norm behaviour. After abnormal
behaviour is detected, the analyst wants to find the causes, i.e. the set of financial variables respon-
sible. The traditional methods frequently used in accounting are variance decomposition and analysis
of ratios in a Du Pont model (Fridson and Alvarez, 2002). Today’s information systems for auto-
mated financial diagnosis and IFC have little explanation or diagnostic capabilities. Such function-
ality can be provided by extending these systems with an explanation formalism, which supports
the work of human analysts in diagnostic processes. In this paper, we describe how the diagnostic
process is fully automated and implemented in a computer program to support decision-makers. It
is applicable to all kinds of underlying business models consisting of identities and behavioural
equations, with the Du Pont and, for example, OnLine Analytical Processing (OLAP) business
databases as special cases.
Diagnosis is generally defined as finding the best explanation of observed symptoms of a system
under study. This definition assumes that we know which behaviour we may expect from a correctly
working system. Diagnosis of business performance is defined in Feelders (1993) as explaining the
difference between the actual performance of a company and its norm performance. The norm