PRIVATE FIRM VALUATION: MAXIMIZING THE ACCURACY OF
SYNTHETIC MULTIPLES BY SELECTING THE OPTIMAL PEER
POOL
Stefan O. Grbenic
Graz University of Technology, Institute for Business Economics and Industrial Sociology, Department of
Management Control, Accounting and Finance, Kopernikusgasse 24/II, A-8010 Graz, Austria
Management Control, Accounting and Finance, Kopernikusgasse 24/II, A-8010 Graz, Austria
DOI: https://doi.org/10.31410/limen.2018.648
4th International Scientific β Business Conference LIMEN 2018 β Leadership & Management: Integrated Politics of Research and Innovations, Belgrade β Serbia, December 13, 2018, CONFERENCE PROCEEDINGS published by: Association of Economists and Managers of the Balkans, Belgrade, Serbia; Faculty of Engineering Management β Belgrade, Serbia; Modern Business School β Belgrade, Serbia; Faculty of Business and Management Sciences β Novo Mesto, Slovenia; Faculty of Business Management and Informatics β Novo Mesto, Slovenia; Business Academy Smilevski β BAS, Skopje, Macedonia; BAS Institute of Management, Bitola, Macedonia, ISBN 978-86-80194-15-8
Abstract
In valuing private firms using transaction multiples, peer transactions must be
identified to generate the synthetic multiple forecast for the firm being valued. Among others,
questions arise concerning the definition of the pool of which the peer transactions must be
selected from. Generally, based on a regional categorization, the peer pool could be either
defined in a broad way consisting of all countries taken into consideration or it could be defined
narrowly consisting of countries with certain comparable transaction-related impact factors.
Finding the βbestβ peer pool follows the basic assumption that the selection of the βbestβ peers
leads to a higher accuracy of the synthetic multiple and lowers the valuation error of the
forecast.
Until date, to my best knowledge, minor related research only exists concerning public firm
valuation. As it is widely accepted that the market for private firms differs from the public stock
market and therefore valuation procedures applicable to the public stock market cannot be used
directly and unchanged in valuing private firms, in line with redefining stock market peer
selection rules, also the rules on peer pooling have to be adapted.
This paper investigates the procedure to optimally define the peer pool in valuing European
private firms. The time period taken into consideration is January 1, 2007 as to December 31,
2016. The transaction data and the related firm data is extracted from ZEPHYR database and
ORBIS database. First, I investigate which regionally defined peer pool the peers of firms in
different countries should optimally be selected from and if the use of a more accurate peer
pool significantly increases the accuracy of the synthetic multiple forecast. Second, I investigate
if a closer definition of the peer pool (according to the country GDP, the country specific GDP
per capita as well as simply the size of the market measured by the number of transactions
recorded) significantly increases the accuracy of the synthetic multiple forecast. Finally, I
investigate the existence of statistically significant differences between major industries.
All computations are based on the roll-out procedure to define the peer group, minimizing
several accepted homogeneity measures (scaled prediction error, scaled absolute prediction
error, scaled squared prediction error as well as the scaled natural logarithm prediction error)
in combination with bias and error minimizing compression methods (arithmetic mean,
harmonic mean, geometric mean, median as well as value-weighted mean). The statistical
significance is measured using a nonparametric test procedure. To avoid distortions caused by
the use of different multiples, all computations are based on the aggregated results of all major
equity and entity multiples.
β
identified to generate the synthetic multiple forecast for the firm being valued. Among others,
questions arise concerning the definition of the pool of which the peer transactions must be
selected from. Generally, based on a regional categorization, the peer pool could be either
defined in a broad way consisting of all countries taken into consideration or it could be defined
narrowly consisting of countries with certain comparable transaction-related impact factors.
Finding the βbestβ peer pool follows the basic assumption that the selection of the βbestβ peers
leads to a higher accuracy of the synthetic multiple and lowers the valuation error of the
forecast.
Until date, to my best knowledge, minor related research only exists concerning public firm
valuation. As it is widely accepted that the market for private firms differs from the public stock
market and therefore valuation procedures applicable to the public stock market cannot be used
directly and unchanged in valuing private firms, in line with redefining stock market peer
selection rules, also the rules on peer pooling have to be adapted.
This paper investigates the procedure to optimally define the peer pool in valuing European
private firms. The time period taken into consideration is January 1, 2007 as to December 31,
2016. The transaction data and the related firm data is extracted from ZEPHYR database and
ORBIS database. First, I investigate which regionally defined peer pool the peers of firms in
different countries should optimally be selected from and if the use of a more accurate peer
pool significantly increases the accuracy of the synthetic multiple forecast. Second, I investigate
if a closer definition of the peer pool (according to the country GDP, the country specific GDP
per capita as well as simply the size of the market measured by the number of transactions
recorded) significantly increases the accuracy of the synthetic multiple forecast. Finally, I
investigate the existence of statistically significant differences between major industries.
All computations are based on the roll-out procedure to define the peer group, minimizing
several accepted homogeneity measures (scaled prediction error, scaled absolute prediction
error, scaled squared prediction error as well as the scaled natural logarithm prediction error)
in combination with bias and error minimizing compression methods (arithmetic mean,
harmonic mean, geometric mean, median as well as value-weighted mean). The statistical
significance is measured using a nonparametric test procedure. To avoid distortions caused by
the use of different multiples, all computations are based on the aggregated results of all major
equity and entity multiples.
β
Key words
Peer Pooling, Peer Group, Business Valuation, Private Firm Valuation, Market
Approach
β
Approach
β
