Forecasting recoveries in debt collection: Debt collectors and information production

Published date01 June 2020
Date01 June 2020
DOIhttp://doi.org/10.1111/eufm.12242
© 2019 The Authors. European Financial Management published by John Wiley & Sons Ltd.
Eur Financ Manag. 2020;26:537559. wileyonlinelibrary.com/journal/eufm
|
537
DOI: 10.1111/eufm.12242
ORIGINAL ARTICLE
Forecasting recoveries in debt collection: Debt
collectors and information production
Johannes Kriebel
1
|
Kevin Yam
2
1
Finance Center Münster, University of
Münster, Münster, Germany
2
Seghorn AG, Bremen, Germany
Correspondence
Johannes Kriebel, Universitätsstraße
1416, 48143 Münster, Germany.
Email: johannes.kriebel@wiwi.uni-
muenster.de
Abstract
Recent theoretical work suggests that debt collection
agencies play an important role in gathering and
processing debtor information. We study a comprehen-
sive data set with information provided by original
creditors and information gathered in thirdparty debt
collection. In line with the theoretical results, the initial
information is sparse and the gathered information is
essential for betterinformed predictions.
KEYWORDS
debt collection, loss given default, recovery rate
JEL CLASSIFICATION
G3; G21; G22; G29
1
|
INTRODUCTION
The management of accounts receivable and accounts payable plays a vital role in the balance
sheets of many producers of goods and services. It is common practice in many industries (e.g.
insurance, telecommunications, and mail order services) to commission specialized collection
EUROPEAN
FINANCIAL MANAGEMENT
----------------------------------------------------------------------------
This is an open access article under the terms of the Creative Commons AttributionNonCommercialNoDerivs License, which permits
use and distribution in any medium, provided the original work is properly cited, the use is noncommercial and no modifications or
adaptations are made.
We are deeply indebted to Seghorn AG for providing the data set and kindly giving a wide range of advice on the debt collection
industry that enabled this study. We thank John Doukas and an anonymous referee for valuable input and suggestions. We are
further grateful to the participants and discussants of the HypoVereinsbank Seminar 2017 in Bochum, Germany; the German
Operations Research Society Financial Management and Financial Institutions Workshop 2017 in Magdeburg, Germany; the
European Conference on Data Analysis 2017 in Wroclaw, Poland; the 35th Annual Conference of the French Finance
Association in Paris, France; the 2018 Annual Meetings of the European Financial Management Association in Milan, Italy;
and Christopher Jung, Antonio Della Bina, Henning Cordes, Jörn Debener, Judith Schneider, and other members of the
Finance Center Münster for their valuable comments, which helped improve this study.
agencies to collect distressed receivables. Similarly, banks tend to resort to collection agencies in
difficult cases (Thomas, Matuszyk, & Moore, 2012). According to industry studies, collection
firms managed a total of 60 billion of receivables in Germany at the end of 2015 (Bülow, 2016)
and over $792 billion in the United States at the end of 2016 (Ernst & Young, 2017).
Surprisingly, little is known about how collection agencies manage accounts successfully.
There is, in particular, little knowledge about what factors drive collection recoveries. Thomas
et al. (2012) and Han & Jang (2013) have studied these factors for bank loans, Beck, Grunert,
Neus, and Walter (2017) have studied them for other goods and services, and Hoechstoetter,
Nazemi, Rachev, and Bozic (2012) have examined the choice of prediction models. To the best
of our knowledge, these are the only results in this field.
A collection agency can use information from different sources to predict collection recoveries.
One important distinction is between the information original creditors provide and the
information debt collection agencies gather themselves.
1
Earlier empirical work, such as that of
Hoechstoetter et al. (2012), Thomas et al. (2012), and Beck et al. (2017), mainly focuses on initially
available information. Thomas et al. (2012) and Beck et al. (2017) find this information to be sparse.
Interestingly, recent theoretical work, such as that of Drozd & SerranoPadial (2017), suggests
that debt collection plays an important role in gathering and processing debtor information.
Fedaseyeu & Hunt (2018) place stronger emphasis on reputational issues but note that debt
collection agencies in particular gain information on a customers willingness to pay. There is no
empirical evidence on the importance of informationproductionindebtcollectionagenciestodate.
Given these considerations, the main purpose of this study is to examine how valuable
information gathered in thirdparty debt collection is for collection recovery prediction. We make
three important contributions to the literature. First, we study predictive characteristics with a large
proprietary data set of more than 300,000 distressed claims in a field where there is no publicly
available data and almost no academic literature to date. This allows our study to complement the
few earlier works on the drivers of collection recovery. Second, we explicitly apply methods of
calculating variable importance to the predictive characteristics. We thus show that the initially
provided information is not only sparse but also insufficient for making precise predictions. Third
and this is our most important contributiondebt collection agencies make use of a much more
valuable set of information that is gathered from external sources, repeated contact with the debtor,
investigations into a debtors financial situation, and experience with a debtor in general.
The remainder of this paper is structured as follows. In Section 2, we briefly review the
available literature on debt collection agencies. Section 3 introduces our data set and descriptive
statistics. Section 4 presents the research design and the regression results. Section 5 describes
our various robustness checks. Section 6 concludes.
2
|
RELATED LITERATURE
There is generally little theoretical and empirical work on the debt collection industry. Beck et al.
(2017) study predictive characteristics of collection on a German debt collection data set. They find
the average level of collection to be around 65%, with a strong bimodal shape. Thomas et al. (2012)
study the differences between inhouse and thirdparty collection claims on loans issued in the
1
We refer to the original creditors collection process as inhouse collection and to the later debt collectors process as thirdparty collection. We use the term
debtor when referring to the individual or company that is overdue on one or more claims. The initial holder of the claim is referred to as the original creditor or
business partner.
538
|
EUROPEAN
FINANCIAL MANAGEMENT
KRIEBEL AND YAM

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT