2nd International Summer School on Intellectual Capital 22

IC-based value creation process of firms: cluster approach
Grigorii Teplykh
Marina Oskolkova
The research is carried out in the framework of "Science Foundation HSE" program,
grant № 13-05-0021
2nd International Summer School on Intellectual Capital
Motivation for study
• A lot of studies is devoted to analysis of Intellectual
capital transformation to corporate value.
• Researchers underline the importance of an analysis of
separate firm.
• Each company has own model of transformation of
intangible assets into performance.
• However an individual specific of firms is usually
ignored or simplified.
• This study is an attempt to look more deeply into
heterogeneity of value creation process and reveal
dissimilarities between different kinds of firms.
2nd International Summer School on Intellectual Capital
Contradiction
• Process of
corporate value
creation is very
individual
Vs.
• Econometric
analysis of value
creation is based
on sufficiently
large set of firms
Any empirical investigation is a compromise – we
analyse common properties of large set of very
heterogeneous firms
2nd International Summer School on Intellectual Capital
Fight with
Heterogeneity
• There are several ways to treat with firm
heterogeneity in studies of the IC topic:
– imposing restrictions on the sample to ensure
needed level of firms homogeneity;
– introducing companies special features into
regression equation as control variables;
– dividing data into the sub-samples and analyzing
each of them separately.
2nd International Summer School on Intellectual Capital
Imposing restrictions
• Companies of one industry, country, firms of
comparable size etc.
– Simplest way to make comparable sample
– Problem of choosing adequate restrictions
– Differences in sample restrictions employed in
different research make their results incomparable
– Chosen restrictions may lead to substantial shortening
of the sample
– There could be sample unrepresentativeness and
impossibility to spread obtained results to companies
beyond the sample
2nd International Summer School on Intellectual Capital
Control variables
• Including control variables (size, ownership,
industry dummies) into value creation model
– Assumption that factors influence the performance
indicator as shift
– The approach doesn’t imply hard sample restriction.
– Strong assumption about independence between
control variables and other tangible and intangible
factors that are included into the value model.
– The correct model that reflects interaction between
independent value factors should be chosen. If their
interaction is ignored then regression estimation is
going to be biased and therefore distorts
understanding of corporate growth drivers.
2nd International Summer School on Intellectual Capital
Sub-samples
• Dividing of a sample into a more homogeneous
sub-samples and analyze the value creation
process in each of them separately
– Unusual way (Youndt et al., 2004; Cheng, 2004
implemented cluster analysis).
– Clusterisation should be made on the base of some
distinguishing indicators, that may affect the value
creation process. So it’s necessary to justify these
variables and boundaries of firm’s group.
2nd International Summer School on Intellectual Capital
The key idea of the study
• Current study aims to reveal companies’ clusters and
investigate IC-based value creation process for each of
them. It is supposed that some factors, such as size and
activity of IC implementation influence the process of IC
transformation into firm value.
• The authors expect to receive different results for
different clusters that are tied with importance and
significance of particular IC assets:
– H1. Size of companies influences value creation model.
– H2. Companies’ IC size influences value creation model.
• Until the gap is closed researchers will may regularly face
with great discrepancy of findings about relationships
between IC and corporate value.
2nd International Summer School on Intellectual Capital
Methodology
• Cluster analysis (k-mean method)
– Grouping of firms into comparable sets
• Firm size (market capitalization)
• Intangible capital involvement (IA book value)
• Factor analysis (PCA)
– Accounting integral indices for each group of
intellectual assets (human, structural and relational
capitals) at the base of proxy’ set
• Regression analysis (2SLS)
– The endogeneity problem is taken into consideration
– Modeling of value creation
• For all firms and separately for each cluster
• For all periods and two sub-periods apart
2nd International Summer School on Intellectual Capital
Final centers of clusters
Final centers of clusters
Cluster 1 Cluster 2 Cluster 3 Cluster 4
Variable
The whole sample
Average logarithm of market capitalization 14,617
11,5356
13,2462
11,2707
Average Intangible to fixed assets ratio
13,3299
9,8438
11,3345
7,3225
Number of companies in cluster
36
63
79
38
Variable
Before the crisis (2005-2007)
Average logarithm of market capitalization 14,853
11,815
13,467
11,356
Average Intangible to fixed assets ratio
13,216
9,753
11,114
7,361
Variable
During the crisis (2008-2009)
Average logarithm of market capitalization 14,286
11,091
12,972
11,091
Average Intangible to fixed assets ratio
13,469
9,944
11,561
7,287
•
•
•
•
Cluster 1 – large-size companies with greater involvement of IC;
Cluster 2 – middle-size companies with greater involvement of IC;
Cluster 3 – large-size companies with less involvement of IC;
Cluster 4 – middle-size companies with less involvement of IC.
2nd International Summer School on Intellectual Capital
Indexes for IC assets (loadings
of first principal components)
HC Index
Variable
Loadings
Board
of 0,2839
directors’
qualification
Share
of -0,2745
directors that
own company’
shares
Personnel costs 0,6473
per employee
Employee
0,6520
efficiency
SC Index
RC Index
Variable
Loadings Variable
Loadings
Patents
0,5728
Company’s age -0,0435
ERP systems 0,5079
implementation
Well-known
brand
0,2931
Strategy
0,4421
implementation
Intangible
-0,1550
assets
R&D expences 0,4410
Citations
in
search engines
Corporate site
quality
Location in the
capital
Location in the
megapolis
0,5288
0,3757
0,4476
0,5395
2nd International Summer School on Intellectual Capital
Model
MCapt  A  BVt  NEt  exp
a1
a2
a3HCt a4 SCt a5 RCt a6 DCt a7 DIt 
log( MCapt )  a0  a1 log( BVt )  a2 log( NEt )  a3 DCt  a4 DI t 
 a5 HCt  a6 SCt  a7 RCt  
•
•
•
•
•
•
•
MCap is market capitalization;
BV is company’s assets book value;
NE is company’s number of employee;
HC is human capital index;
SC is structural capital index;
RC is relational capital index;
DC – dummy variable that has the value “1” if the company operates in
Great Britain and “0” otherwise;
• DI – dummy for industry (1 if service or 0 if manufacturing)
• A, a1…a7 are regression coefficients; t is a time period; ɛ is error.
2nd International Summer School on Intellectual Capital
Regressions results
(all periods)
Variable
C
Log (BV)
Log (NE)
Country dummy
Industry dummy
HC Index
SC Index
RC Index
Cross-sections obs.
Total panel obs.
Adjusted R-squared
All companies Cluster 1 Cluster 2 Cluster 3
the whole sample
Model 1
Model 2 Model 3 Model 4
1,578***
7,685*** 5,704*** 5,659***
0,819***
0,511*** 0,341*** 0,645***
0,012
-0,126
0,241** -0,159**
0,332***
-0,025
-0,023
0,030
0,084
0,494***
-0,042
0,015
0,088*
0,028
0,073
-0,007
0,115***
0,046
0,012
0,133***
0,146***
0,209*** 0,163***
0,009
212
36
61
78
771
136
223
283
69,5%
53,8%
39,7%
41,4%
Cluster 4
Model 5
6,053***
0,555***
-0,166
0,367
-0,166
0,199
0,433***
0,219*
37
129
36,7%
2nd International Summer School on Intellectual Capital
Regressions results
before the crisis (2005-2007)
Variable
C
Log (BV)
Log (NE)
Country dummy
Industry dummy
HC Index
SC Index
RC Index
Cross-sections obs.
Total panel obs.
Adjusted R-squared
All companies Cluster 1 Cluster 2 Cluster 3 Cluster 4
before the crisis (2005-2007)
Model 6
Model 7 Model 8 Model 9 Model 10
1,412***
5,994*** 6,375*** 5,630*** 5,707**
0,884***
0,638*** 0,298** 0,689***
0,608*
-0,033
-0,082
0,254*
-0,180**
-0,206
0,34***7
-0,042
0,086
-0,019
0,542
0,041
0,301**
0,066
-0,074
-0,043
0,054
0,001
0,087
-0,059
0,175
0,100**
0,041
0,019
0,115**
0,378*
0,136***
0,122**
0,081
0,053
0,184
204
34
60
73
37
387
66
115
138
68
71,2%
64,6%
33,1%
45,4%
24,6%
2nd International Summer School on Intellectual Capital
Regressions results
during the crisis (2008-2009)
Variable
C
Log (BV)
Log (NE)
Country dummy
Industry dummy
HC Index
SC Index
RC Index
Cross-sections obs.
Total panel obs.
Adjusted R-squared
All companies Cluster 1 Cluster 2 Cluster 3 Cluster 4
during the crisis (2008-2009)
Model 11
Model 12 Model 13 Model 14 Model 15
1,660***
8,902*** 4,796** 5,697*** 6,915**
0,771***
0,467*** 0,400** 0,604***
0,464*
0,040
-0,249
0,232
-0,139
-0,129
0,319***
-0,039
-0,119
0,096
0,141
0,123
0,744***
-0,185
0,088
-0,306
0,110
-0,008
0,048
0,054
0,223
0,134**
0,045
0,017
0,154**
0,489**
0,155***
0,316**
0,259**
-0,053
0,283
204
36
59
77
32
384
70
108
145
61
66,2%
40,3%
25,1%
28,1%
40,2%
2nd International Summer School on Intellectual Capital
Results
• Importance of tangible resources (labour and
capital) declined and the significance of
intellectual resources (HC, SC, RC) increased
during the crisis – for all the clusters
• Different value creation model for each cluster
either before and after the 2008 crisis
• Different response for each groups of firms to the
2008 crisis
• These conclusions are just visual. We have not yet
tested models for structural shits.
2nd International Summer School on Intellectual Capital
Conclusions
• We proposed a cluster approach to analysis of
firm heterogeneity in value creation model
• We revealed that firm heterogeneity is more
important that it was assumed earlier
• We found the value creation depends on such
factors as firm size and amount of intangibles.
– They impact on effects of value drivers (both
tangible and intellectual assets)
– They impact on change of effects of tangible and
intellectual assets during the crisis