PreprintPDF Available

Organizational Values and Firm Profitability: A Financial and Cultural Perspective on Europe

Authors:
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

The following paper develops a linkage between organizational cultural values and firm profitability. Reviewing current research, it becomes clear in analysis regarding firms profitability, culture played only a minor side role. However, scholars stated the role of cultural values on organizational development which became a core motivation for this paper. We build on current research that stated a strong linkage between firm performance and organizational culture. The construction of the list of profit enhancing organizational values is based on a data set of ca. 4.000 value statements espoused by ca. 150 organizations in Europe. The values are linked to three different profitability measures which are derived from current research. We analyze which of the values have the highest impact on firm profitability and with that characterize a profitable firms organizational culture. The results show a domination of business values that are customer focused and product related, furthermore values related to human resource play a significant role. By advancing our understanding of organizational cultural values and its linkage to firms profitability, we provide a means by which future research into organizational values and corporate culture can progress.
Content may be subject to copyright.
Organizational Values and Firm Profitability: A
Financial and Cultural Perspective on Europe
BJOERN PREUSS
ABSTRACT
The following paper develops a linkage between organizational cultural values and
firm profitability. Reviewing current research, it becomes clear in analysis regarding
firms profitability, culture played only a minor side role. However, scholars stated the
role of cultural values on organizational development which became a core motivation
for this paper. We build on current research that stated a strong linkage between firm
performance and organizational culture. The construction of the list of profit enhancing
organizational values is based on a data set of ca. 4.000 value statements espoused by
ca. 150 organizations in Europe. The values are linked to three different profitability
measures which are derived from current research. We analyze which of the values
have the highest impact on firm profitability and with that characterize a profitable
firms organizational culture. The results show a domination of business values that are
customer focused and product related, furthermore values related to human resource
play a significant role. By advancing our understanding of organizational cultural
values and its linkage to firms profitability, we provide a means by which future research
into organizational values and corporate culture can progress.
Mr. Preuss is with the Department of International Economics and Management at Copenhagen
Business School and with the Department of Economics at Radboud University Nijmegen
I. Cultures’ Impact on Profitability
Recent paper have showed that values and and belies have an impact on organiza-
tions. Papers mapped out, how values look in organizations. Those values can also be
used to explain organizational development and actions. (Rokeach, 1968) According
to current research corporate values played an important role when it comes to the
influence on organizational development and performance. (Kabanoff, Waldersee, and
Cohen, 1995) Internal values and cultural set ups of the organizations are discussed
widely among managers together with their impact on performance and the develop-
ment of the organization. (Bansal, 2003) Recent studies have linked the values and
beliefs of organizations to their performance like Calori and Sarnin (1991), the in-
tegration of social norms (Grøgaard and Colman, 2016) and the commitment of the
organization. (Howell, Kirk-Brown, and Cooper, 2012)
For the measurement of culture, different methodological approaches have been
chosen. (Schwartz, 1992) Most of the widely used approaches are focused on na-
tional culture and the information is retrieved by employing questionnaires. Those
methods have certain limitations that influence the accuracy, bias, cost and time of
studies. Recent studies in the field of organizational values and culture have either
performed questionnaires or mapped culture based on textual statements from the cor-
porate homepage of organizations. (Hofstede, 1983) (Bourne, Jenkins, and Parry,
2017) Through out the possibilities to collect vast amounts of that, research on how
to make use of unstructured and textual data with quantitative methods has increased
in recent years. (Bourne et al., 2017) The published information about organizational
cultural values though web platforms like ”Glassdoor” (www.glassdoor.com) and ”In-
deed” (www.indeed.com) present employees’ values and descriptions of a companies
work environment.
2
Impact and the relation of organizational values to economic performance should
be subject of more academic studies. Therefore the aim of this work is to map or-
ganizational values and culture that has an impact on firms economic performance.
Those values are derived from what employees actually experience instead of official
published values like in recent research. (Bourne et al., 2017) It will build up a ranking
of values and cultural concepts that represents the value exposure of well and not so
well performing organizations in Europe. The leadership culture and its related val-
ues, together with its synonyms, is grounded in the mind of many employees across
countries and so it is sufficiently comprehensive to capture value terms relevant to a
variety of contexts. We present as an outcome of this paper a structure of organiza-
tional values that distinguish economic well performing organizations from those who
do not. The study does this empirical through out the analysis of ca. 4.000 judgments
of organizational values in major European stock market listed companies. The out-
come of the application of supervised machine learning algorithms reveals a structural
arrangement of organizational cultural values linked to profitable firms, that allows us
to compare this with other structural arrangements in the organizational culture litera-
ture. With this we can consider the wider implications for leadership and management
of organizations that aim to increase their economic outcome.
The paper is structured as follows. First we review the current literature around
organizational culture and the measurement of economic performance. Furthermore,
we develop a methodological framework for analyzing the statements of organizational
cultural values and a possibility to analyze the predictive power of certain exposed
values towards economic performance. We summarize the empirical findings and from
there we place the study in the context of contemporary studies in organizational value
and cultural studies. In the end we aim to give considerations on how the mapping of
organizational cultural values can help drive management and leadership.
3
II. Organizational Cultural Values
According to a common definition among researchers and practitioners, culture is
seen as a system of values, beliefs, and assumptions that are shared among employees.
(Hofstede, 1983) The culture is to a large extend influencing how an organization is
managed. (Rosenblatt, 2011) (Migliore, 2011) (Weber, Shenkar, and Raveh, 1996)
In current research, cultural frameworks such as the one of Hofstede and House et
all. are dominating the field of research. (Hofstede, 1983) (House, Javidan, Hanges,
and Dorfman, 2002) Both started out with the focus on the country specific concept of
culture. House et all. had from the beginning an organizational focus whereby Hofstede
later added organization focused concepts. As like House, they build the connection
between the organizational culture and the one of countries such as the GLOBE study
from House. (House et al., 2002) These studies that started with he country based
understanding of culture tried to expand their method and framework to organizations
which was claimed from Hofstede as wrong or limited. (Hofstede, 1983) According to
Hofstede, organizational culture needs to be measured in a different way in order to
capture the right values and believes that drive the organization.
The focus on specific values in organizations has been the focus of papers from
researchers that aimed to describe it in more detail on a corporate level. (Gagliardi,
1986) (Schein, 2010) (Calori and Sarnin, 1991) The impact that the culture has
on organizations has been shown by researchers in recent times. (Warner-Søderholm,
2012) Studies focusing on specific corporate processes such as Mergers and Acquisi-
tions presented also a relations between the organizational cultural set up and the
profit derived from the transaction. (Janson, 1994) However, those studies do not elab-
orate on specific cultural elements but the differences in cultural set ups. This leaves
most of the organizational culture in the dark and does not defines specific cultural
4
values that have an impact on the organizations. However, critics of those approaches
claim that those might not be appropriate to capture the complete disperse field of cul-
tural values. (Schwartz, 1992) The recent approach has the advantage that compared
to the classical frameworks it is purely derived from exposed organizational values.
It is designed to capture nuances of culture specifically in organizations much more
in detail then the frameworks that are designed for country wide culture. (Hofstede,
Neuijen, Ohayv, and Sanders, 1990) When it comes to organizational values, one of the
best-known framework of organizational values is Quinn and Rohrbaughs ”Competing
Values Framework”. (Quinn and Rohrbaugh, 1983) This incorporates two main com-
peting dimensions: organizational focus (internal vs. external) and preferred structure
(stability and control vs. flexibility and change). These dimensions can be related
to four types of organizational effectiveness: human relations; open systems; internal
process; and rational goal. All types can be connected to a set of values. Quinn and
Rohrbaugh did not define those so that other have developed measures of organizational
values. (Bluedorn, Kalliath, Strube, and Martin, 1999) (Ostroff, Shin, and Kinicki,
2005) Researchers such as Calori and Sarnin (1991) used a framework containing ques-
tionnaires and textual analysis to describe culture in organizations. There sample was
very small with just a couple of organizations but they started to develop a framework
that enables to research the connection between culture and firms profitability.
Building on the limitations that classical cultural frameworks had, other work has
focused more on comparing values communicated from management of organizations.
Recent technological development together with the exposure of more and more data
on the internet, enabled a different kind of research approach than questionnaires.
They focused on the values of person working in an organization to derive an answer
on personal organizational fit. (Cable and Edwards, 2004) (Kristof, 1996) With this
the position of Hofstede regarding a specific measure of culture on an organizational
5
level, got supported by other researchers in the field. Schwarz added to this discussion
that organizational values are different from, but related to, individual, cultural and
societal values. (Schwartz, 1992) (Calori and Sarnin, 1991) Calori constructed in his
wok a framework that was specifically designed to link the organizations culture with
firm performance. In his research, Calori presented a mixed method approach that to
some extend utilized the exposed values in textual form. Calori and Sarnin (1991)
In contrast to this stands House with his GLOBE study that places both in the same
framework. (House et al., 2002)
III. The Impact of Culture on Firm Profitability
Culture impacts society and with this also companies in the society as well as
people working in those companies. According to Kreps (1990), culture matters in
an organizational context since there are factors that cannot be regulated ex ante.
Only few scholars have carried out studies on the relation between culture and eco-
nomic performance of firms. This was done mainly by frameworks which are based
on questionnaires and scales to measure a form of organizational value, whereas some
researchers use measures across value forms by applying statistical methods. (Calori
and Sarnin, 1991) The role that culture both on a national as well as organizational
level plays in an organizational context and its relation to its economic situation was
researched by Guiso, Sapienza, and Zingales (2015). Further work in this area showed
evidence that culture is a core driver of value creation when it comes to specific corpo-
rate situations such as restructurings or mergers and acquisitions which aim to increase
the companys profitability. (Sarala and Vaara, 2010) The degree that culture influ-
ences organizations market orientation and with this profitability differs from country
to country. (Selnes, Jaworski, and Kohli, 1996)
6
The different roles that culture plays is, as Selnes et al. (1996) states, not only
depended on the country but also the industry which might develop own working
cultures. Calori and Sarnin (1991) All industries have their own organizational culture
or value set up that organizations tend to follow. Edmans (2011) showed in his research
that companies that were ranked high for good working conditions also had a significant
increase in stock price return. However, research from Guiso et al. (2015) has shown
that not only culture impacts organizational profitability but also the financial decisions
that are made in companies impact the culture in those. Stating those lets us think
about a cause and effect problem which is in current research not clear answered.
However, even though cultural values and norms can be value enhancing they often
are also linked to costs which are a loss in flexibility if values are defined too strict or
extreme.
But which value set up is one that enhances profitability and economic prosperity of
an organization and which is harmful to it. The study from Guiso et al. (2015) showed
that variables such as hard working, quality, integrity and teamwork among other had
an impact on organizational profitability. Out of their event study of IPOs and their
stock market return they concluded that integrity seems to have the biggest impact
of all the listed values. The influence that cultural values have on the motivation of
employees and with this link to economic profitability of a firm was proven by current
research. (Gruman and Saks, 2011) That some the from House, Hanges, Javidan,
Dorfman, and Gupta (2004) defined cultural norms such as performance orientation or
future orientation have a positive impact on the economy of firms is straight forward.
But how important the role of those is, is still not entirely clear researched. This ques-
tion will be the connecting point to the following research on organizational cultural
statements and their relation to firm profitability. But first we look at the question
what defines a firm as profitable or not profitable and how it can be measured.
7
IV. Measurement of Organizational Profitability
The measurement of firms performance is among academics a widely discussed field
and this even with neglecting the philosophical discussion on what a firms performance
should be. Recent interdisciplinary research between culture and finance such as the
one from Calori and Sarnin (1991) used the measures ”return of assets”, ”return on
sales” and ”growth in turnover”. From the economic perspective that is followed here,
two different approaches of financial performance exist. One group is oriented on a
firms stock price which should according to the efficient market hypothesis incorporate
all relevant economic information of the firm. (Fama, 1970) (Fama, 1991) The other
group focuses on fundamental valuation approaches using ratios and enterprise value
models. (Carlisle, 2014) (Loughran and Wellman, 2011) The point of view of the later
approach is that a company can have value that is different from the one the market
assigns to it. One of the first scholars following this view was Benjamin Graham
who wrote the famous book ”The intelligent investor” which is one of the first work
related on valuation of securities. (Graham and McGowan, 2005) In this study we
follow the later view that bases the definition of companies success and profitability
on fundamental measurements in the manner of Graham. (Graham and McGowan,
2005) Much recent research updated and tested Grahams approach. Especially the
selection of the most meaning full ratios was core of many papers in the field such
as: Piotroski (2000), Aggarwal, Hiraki, and Rao (1992) and other. The aim was
to find the one valuation metric that can differentiate a profitable company from an
unprofitable. (Carlisle, 2014)
In recent literature, three ratios have been tested to define profitable companies. (Gray
and Carlisle, 2012) (Carlisle, 2014) In some cases those have also been combined to
give a more comprehensive picture of a company when selecting it with such a mea-
8
surement tool. (Carlisle, 2014) All measurements have in common that the set a value
measurement in relation to a return measurement. (Basu, 1977) The ratios that have
been selected for this research are ROA,ROE and EB IT DA/Enterprise value. This
shows that with the ROA measure the paper links up with the methodological decision
of Calori and Sarnin (1991) but also extended the selection of measures for perfor-
mance evaluation. Research argues in many ways for the different measurements. So
is the ROE often used in investment research. It defines how much money a company
makes with the used equity. However, it is to some extend difficult to compare different
companies hence their leverage is different which can influence this measure. (Beaver
and Ryan, 2000) As a result of this critic, models where adapted towards the return
on asset ROA. This measure has the full amount of assets as its base an was previous
selected for the purpose of finding evidence in the relation between culture and firm
profitability. (Calori and Sarnin, 1991) Furthermore, this measurement excludes in
contrast to the ROE measure the leverage effect and enables a comparison of different
organizations of different industries which tend to vary in their debt to equity ratio.
The third measure that will be used in this paper is a further development of the return
on equity which also tries to exclude some of the leverage effects. It sets the E BI T DA
in relation to the enterprise value of the firm EV . (Greenblatt, 2010) (Carlisle, 2014)
The enterprise value is hereby defined as the market value of equity plus the value of
debt in the company. (Liu, Nissim, and Thomas, 2002) As with the ROA this model
incorporated equity and debt but as the ROE it lays more focus on the equity part
that is in this case represented with the market value of equity. (Liu et al., 2002)
(Greenblatt, 2010)
9
V. Method and Data
As already mentioned before, parts of the here discussed and used method is based
on previous work of the author on capturing organizational cultural values from text.
(Preuss, 2017) To analyze organizational cultural values, text statements from employ-
ees of large European organizations are used. The data sample contains statements
about organizational cultural values from companies listed at the blue-chip indices
from Finland (OMX Helsinki 25), Norway (OMX Oslo 30), Denmark (OMX Copen-
hagen 20), Sweden (OMX Stockholm 30) and Europe (EuroStoxx 50). The textual
statements contain a description of the working culture in an organization. See for an
example one description of the company SEB AB:
SEB Position: Full time 2 years Pros: Flat hierarchy and the possibility of involve-
ments. The management listens to your ideas and the work environment is friendly.
Good benefits for working. Cons: Lack of internationalization. Small bonuses compared
to others. Too slow development which could lead to a bad position in the global finance
market. Strong focus on the Nordics.”
With a number of ca. 30 of such statements per company it enables us to show
relationships between cultural values represented in the organizations and their prof-
itability. The influential values of each organization give a picture of how employees
see their working culture in the organizations. All statements are to some extend
similar. However, there are certain differences in length and in the information con-
tained. Between the platforms on which employees can upload their descriptions,
there are some differences. Some are more qualitative focused with longer descriptions
(hhtp://www.glassdoor.com/) others are more focused on short precise definitions by
using key words (http://www.indeed.com/) or a combination out of both (Indeed).
Only a small fraction of the statements from both platforms (less then 2 percent) are
10
long and exhaustive. Most descriptions are precise and similar to the presented exam-
ple from SEB AB. The similarity of most of the texts is critical when it comes to the
next step hence it will otherwise lead to a fuzzier data set. Some critic might however
also be addressed to this sort of data. Given the publicly availability, employees of
organizations might fake some of the statements to ensure that the company is pre-
sented in a certain way (too good or too bad). This might to some extend also impact
this research. However, analyzing multiple descriptions statistically might limit the
potential bias.
The second part of the data collection aimed to collect data that helps to describe
a companies profitability and performance from a financial perspective. It therefore
used the before discussed variables and rations to measure whether a company could be
defined as profitable or not. The ratios and their under laying factors are represented
in the following table.
Profitability measure Underlying values
Return on Assets (ROA) Net profit; Total assets
Return on Equity (ROE) Net profit; Book value of equity value
EBITDA/Enterprise Value EBITDA; Debt value; Market value of equity
Table I: Profitability measurements
Those measures were collected and calculated for each company contained in the
sample data set. Each of the profitability measures could then be used as a target
variable of the model. The whole data set can be described by the following table
which summarized the key statistics of the performance measures in the sample. The
following table of summary statistics show that there is a variety in the return measures.
The highest variety can be found at the ROE measure. Looking at the quantiles the
11
differences are lower then by just focusing on the min and max values of each measure.
ROA ROE EBITDA/EV
Min -0.002 -3.130 -0.001
Max 0.528 4.848 0.148
Avg 0.113 0.348 0.055
1.Qtl. 0.060 0.215 0.031
2.Qtl. 0.098 0.277 0.060
3.Qtl. 0.132 0.390 0.077
Table II: Key statistics of profitability measures
By using the company names as a link we were able to link the metrics for economic
performance of each company to the prepossessed and cleaned value statements. By
doing this we received a table including for each statement the value stems as well as
the economic performance of the company to which the value statement belongs to.
A. Text Processing
The data in form of text is traditionally difficult to analyze. Mostly those sort of
information can be used in qualitative studies but not in quantitative research designs.
Using it in a quantitative way and analyze it statistically makes a transformation of
the data necessary. (Preuss, 2017) This is done by a process called text processing.
Text processing uses algorithms that read the textual statements and creates a matrix
out of words and n-grams (combinations of words that occur together). A subset of
the data sets structure after the transformation can be found int he appendix I of this
paper.
12
In the following steps the word matrix can be analyzed with mathematical ap-
proaches to cluster or regress a target value. For this to clean the matrix and transform
it into a unified format is essential. (Preuss, 2017) To improve and clean the content
of the matrix, all words contained in the texts get transformed to make them more
homogeneous. This includes a step to stem the words and bring them back to their
original word stem, a step to eliminate stop words which are irrelevant for the specific
content. The last step is to transform the remaining word stems and n-grams to lower
class meaning that all words are written in small letters (e.g. out of ”He Go high” the
transformation makes ”he go high”). The outcome is a unified matrix of word stems
defining a column and for each original cultural statement there is a row. The matrix
represent then the occurrence of a word stem in a cultural statement.
B. Supervised Regression to Predict Profitability
Getting to know the important word stems that distinguish a profitable organiza-
tions form a non profitable one is the aim of the following step. This process creates
the subset of values that can mostly be associated with firms that have high prof-
itability compared to their peers in the data set. The selected algorithm family for
this is the one of trees. (Lior et al., 2014) Hence the target variables are numbers,
the problem can be defined as a regression problem for which the tree model needs to
be set up. Because we expect a non linear relationship between cultural values and
the profit measurement, we see a linear regression model not capable of explaining
the relation between data points. A tree model however, deals nicely with non linear
relationships and can also display the outcome in a graphical format which supports
in the interpretation. (Loh, 2011)
The algorithm will construct a tree structure that hierarchically orders threshold
rules for the accuracy’s of word stems and n-grams in the data set. Each rule will
13
be linked to one particular word stem and define whether this stem has to be in a
description or not (defined by the signs ¿0.5 meaning accrue and <0.5 not accrue
since the data set is containing 0 and 1)
In general, a decision tree algorithm divides the underlying data set into patterns.
In the case of the research project, the aim is to find the word stems, from the cultural
descriptions, that can be associated with a higher profitability of a firm. The algorithm
searches now for patterns in the data and divides the data items into groups. For this
the algorithm fits threshold rules if a word stems should accrue or not. For example,
the data set will be divided by a group of statements that contain the word stem work.
The algorithm then looks on whether these rules can explain some of the variety of the
target variable (the profitability measure in this case) and later ranks this rules based
on the explanation power of it to construct the tree that can be seen in printed form
in the analysis section of this paper. The way how to interpret the outcome of the
algorithm will be explained along the analysis.
The mathematical principles that are underlying this process, are described in the
appendix three at the end of this paper. The outcome of this approach is for every
target variable (here the profitability measure) a list of most important variables (word
stems), the tree structure together with the threshold rules and a set of plots for the
relative error and the R2. The plots can be found as figure: 6, figure: 4, figure: 5 in
the appendix II of the paper.
C. Sensitivity of applied model
Given this novelty of non linear model in economic research, we applied certain
steps of sensitivity analysis. With these steps we want to build trust in the applied
models and cater for eventual critics as well as be certain about limitations.
The first step was to test the impact of correlated input variables. For this we copied
14
one of the explanatory variables of the model and recomputed it. The result showed no
impact on the model outcome. This was in align with findings from Belgiu and Dr˘agut¸
(2016)). To avoid in our final results any impact of variable bias, we additionally run a
correlations analysis to be sure that the main predictors are not correlated. This was
done to comply with certain discussions around different computation methods for tree
model. Hothorn, Hornik, and Zeileis (2006)
D. Delimitation and Discussion of Approach
The paper applied the anova method to generate the decision tree. This has been
done due to the fact that the target value is a static measure (profitability measure)
and hence this we deal with a regression problem. (Mingers, 1989) The method aims
to get results like an OLS regression solve for a linear relation would be expected. We
assume a non linear relationship and hence this, we use a decision tree algorithm. One
would ask the question why the model only includes textual data describing cultural
organizational values and no other factors which influence according to current research
firm’s profitability (e.g. profitability margin, debt to equity ratio, size, industry, etc.).
The reason why we choose to exclude all other factors is that the method approach
aims to present the predictive power of cultural values on organizational profitability.
Hence the model at no point aimed to explain firm’s profitability completely, including
other variables would bring no more value to answering the research question.
The plots of R2values show that depending on the profitability measurement vari-
able, between 20 and 40 percent of the variety of the profit can be explained by the
occurrence of cultural values, see for this figure: 6, figure: 4, figure: 5. In the tree mod-
els the cultural values (word-stems and n-grams) have been used as variables. This
resulted in contrast to most applied models of current research not in one or two cul-
tural variable but instead in 180 variables that describe the culture of the statements in
15
the organizations. Adding other variables to the model would contribute to increasing
the R2value and with this increase the explanatory power of the model itself, however
would not lead to more understanding of which combination of organizational cultural
values have the highest influence on the profitability of the firm’s.
VI. Outcome and Discussion
The paper sets out to link organizational values with firms profitability and with
this links to the goal of the research of Calori and Sarnin (1991). With the following
analysis, the paper aims to present profitability enhancing organizational values. We
see hereby the values defined by what employees actually stated about their work-
ing culture in the respective organizations. The resulting values help to define value
concepts that are characteristic for profitable organizations.
A. Cultural Values that Define Profitable Organizations
The tree algorithm reviles core concepts that in the best way describe the culture
which can explain the level of profitability in the organizations. The regression based
tree approach of the data lists the following organizational values which have the biggest
impact on the model in a descending order. In the further presented tree map, the
values get in addition to their impact a rule assigned which states whether the impact is
positive or negative. This means that either the presents of a word stem has a positive
effect on the firms profitability or a negative one.
16
Wordstem Model importance
bank 25
product 12
growth 5
sales 5
fast 5
goal 5
help 5
promot 3
custom 3
offic 2
Table III: Tree model importance of top 10 word stems for ROA measure
Looking on those values, the reader might wonder whether all the word stems
actually represent values or if some not more define a particular industry or business
model. Referring to Calori and Sarnin (1991) we want to build on to of the idea that
organizational cultures are also to a large extend influenced by the companies way of
doing business. This means that even industry related stems might carry some specific
industry unique values and can express sth. about the organizational culture of the
underlying organization. The result of compiling the first tree shows that culture that
influence the profitability which is measured by the ROA can be defined by the rank
of values in the following plot.
17
Figure 1. : Map of tree model ROA
The plot shows the hierarchical order of threshold rules for the most important
word stems in that model (here it is the stem ”bank”). There are more rules and word
stems influencing the whole prediction model which are not displayed. The top word
stem is the one with the most explanatory power for the target variable (in this case
the profitability measure ROA). The rule says whether the word stem should occur or
not (this rule says ”bank>0.5” which means that if the stem bank occurs in the data
set, the rule is full filled). When a rule is full filled that means that the tree directs the
reader to the left. On this an other rule can follow until the graph reaches an endpoint
(if in this case ”bank” is in the data, the model predicts a ROA of 0.015). The end
of the arm can either have an end point or a following rule for further differentiation
of cases (the left arm has an end point but if ”bank” was not in the data the rule
gets rejected and the right arm gets followed, here the next rule gets asked). With
18
the n=xthe number of cases is displayed under each node. Reading this tree model,
means that in order to generate a higher ROA, the value ”bank” should play a low role
and if bank occurs in the data set it will define a value statement that is associated
with a lower ROA. This might be connected to previous findings from scholars such
as Edmans (2011) who showed that good working conditions goes hand in hand with
high company values. In banking this was less the case over the last decade and this
might have caused this finding. It might furthermore related to the fact that banks
have a unique way of working which is connected with a lower profitability measure.
This value stem is tightly related to the industry of the company. And when looking
into the sub sample of companies full filling this rule, the data shows that nearly all
companies are banks of finance related companies. This finding can be seen similar to
the findings of Calori and Sarnin (1991) who state that the industry forms same as the
country a unique way of working. Now it has to be said that this finding could also be
interpreted that is is not related to the working but to the fact that banks had in the
time horizon of this study a lower ROA performance than other sectors or industries.
Going to the next, the word stem ”product” as a value should be important, ”goal”
and ”growth” is also important etc. A supportive culture for the ROA measurement
can therefore be defined as a value set that is focused on the ”product”, ”goal” and that
is ”growth” oriented. Filtering the data shows statements with a goal orientation come
in 51 out of 64 cases from Nordic organizations. When adding also product, nearly only
Nordic statements remain with few exceptions. Companies which full fill these rule are
for example Chr. Hansen AS, Coloplast AS, Assa Abloy AB or Sandvik AB which are
mainly Danish or Swedish. On EU level SAP AG aligns with these values 3 times as
well others are only present one time or not at all. When looking on different subsets of
the data where specific values get supported such as ”growth”, also EU organizations
score high, here all sectors are present with the domination of the service sector. Having
19
the product and sales as a core of the organization is much align with the development
of modern management practice who build even their information systems around the
Customer relationship management (CRM) system instead of accounting. This finding
together with the high impact of ”promoting” and ”selling” which also lead to more
ROA show that this common management practice can be supported by these findings.
It can also be related to the findings from Guiso et al. (2015) who presented the finding
that values like hard working ad integrity had the biggest impact on firms profitability.
As a sub conclusion we can say that the values that have a positive impact on the
ROA measure are focused on the core of business. The organizational culture should
be mainly driven by business values which have the product and customer in the center.
Also on a 2nd level product orientation together with customer focused is important
to generate a higher ROA.
To examine a more complete picture the second profitability measure, the ROE, was
analyzed by this study. For this, the next step in the research contained to reproduce
the steps and look whether different values would define companies with high ROE
measure versus those with a low measure. As with the previous step we will look at the
values that are most influential and additionally compare it to the findings we found
for the ROA measure.
20
Wordstem Model importance
bonus 20
grow 16
clear 10
benefit 5
part 5
sale 5
decis 4
field 3
chang 3
employ 3
Table IV: Tree model importance of top 10 word stems for ROE measure
The based on the algorithm produced tree map gives evidence on similarities but
also some differences in enhancing values. The figure 2 shows that values like ”bonus”
should not play a large role which is the case for companies representing the tobacco
and energy business. However, ”benefit” is positive associated with a higher ROE. A
possible interpretation could be that a higher focus on bonus would increase an inher-
ited risk in operations whereby the focus on general benefits could be also qualitative
and with this less risky but still drive employed to higher performance and motivate
them. we can also see that the sub sample of statements including benefit are from
Nordic organizations which according to the literature have a higher focus on the work
environment and non monetary benefits. Values like ”clean” and ”employ” are often
named stems in organizations with a high ROE. A look at the companies complying
with these value stems shows a focus on industries like Pharmacy and other high skill
21
businesses like energy or engineering. Furthermore, those values could inherit trans-
parency and focus on people skills which could refer to the value stem ”employ” . The
focus in values like ”employ” can also be linked to facts like growth with an increase in
employment. Values that would describe a dynamic growth environment like ”change”,
”grow” and ”shift” plays an important and positive role which is also important for the
industries where these values are over represented like the technology industry. This
leads to the interpretation that grow oriented companies would more often generate a
high ROE as well as that a dynamic environment fosters ROE growth in organizations.
Figure 2. : Map of tree model ROE
The analysis shows that for the companies with high ROE, the values are more
employee focused and for the ROA more product based. Especially the employee and
team focus is a major finding of the work from Guiso et al. (2015). The difference
between the two value measures could be because of fundamental differences between
22
companies that report traditionally a higher ROA versus a high ROE. Since both
return measurements can be influenced by the leverage effect the ROE measure would
be higher in higher leveraged companies this would inherit more risk then the more
equity financed companies which eventually score lower in the ROE measure. (Beaver
and Ryan, 2000) The focus on the product and product related values in the companies
that score higher in the ROA is well align with the high product focus in growth
companies in contrast to finance and financial engineering in established companies
where the market is stable and management tend to improve the firms performance
with financial engineering instead of product innovation.
As derived from the literature review, there is a third metric that aims to measure
corporate profitability more objective and less influenced by the leverage effect. This
is the ration between EB IT DA/E V . This ratio focuses at the one side on the market
value of the equity and the E BI T DA value. According to the fact that it excludes the
leverage effect, it should be related to similar values like the ROA measure.
23
Wordstem Model importance
bank 50
offic 4
salari 4
custom 3
clear 3
depart 3
balanc 3
opportun 2
sale 2
life 2
Table V: Tree model importance of top 10 word stems for EBITDA/EV measure
As the values associated with the ROA measure, this list shows a high importance
of ”bank” for the model. The rest of the values are similar to a mix values that
are associated with companies that score high in the ROA and ROE. Both individual
focused values as well as values like ”product” are named and influential. Given the fact
that the measure EB IT D A/EV adds a market value to the perspective of the ROA,
the tree map shows also a subset of important values from market driven companies
and non market driven companies. We can see the relevant values of the EBIT DA/EV
measure as a combination from the other two.
24
Figure 3. : Map of tree model EBITDA/EV
With the defined value set, the study raise some important points. One is that
in contrast to other research on organizational values, those from employees exposed
values are analyzed and not the official ones from an organization. (Gurung and Prater,
2017) The fact that the exposed values are derived from a text mining analysis and
not based on questionnaires, separates the study more from the main work of other
scholars in this field. (Gurung and Prater, 2017) However, approaching the question
towards organizational values with this enables the researcher to draw conclusions on
a organizational level. (Gurung and Prater, 2017) Having this opportunity, it fosters
analysis steps regarding a specific definition of a supportive or enhancing cultural set
up for various targets such as firm profitability. (Calori and Sarnin, 1991) In contrast
to recent research that looked on just the impact of culture as one factor among other
on a firm profitability, this study analyzes the element of culture in much more detail.
25
this is due to the fact that culture and its elements are at the core focus of this paper
as it has been done on a questionnaire basis from Calori and Sarnin (1991). Having
said this we can, by using the tree based algorithm, define enhancing culture for firm
profitability. This is done by interpreting the core values that are characteristic for the
cultural subset that we associated with profitable companies.
A necessity to define limiting cultural values or the one that impacts profitability in
a negative way is form our point of view already incorporate in the approach. This view
is derived from the fact that with the tree based model we define the core differentiation
between organizations with high and low profitability. Having done this means that
for organizations having a lack in these derived values would be found in organizations
with lower profitability. It add on top of research from Guiso et al. (2015) who
presented a value set that overlaps to a large extend with the result of this study.
This values can be linked to firm profitability. In conclusion of the analysis it can be
stated that organizational cultural values impact the profitability of an organization.
Analysis showed that between 20 to 40 percent of the profitability of organizations can
be explained by its organizational cultural values, see figure 4, 5 and 6 in the appendix.
With this relevance the study supports Kreps (1990), Sarala and Vaara (2010) and
Guiso et al. (2015) who stated that culture matters in an organizational context and
influences an organizations outcome.
VII. Conclusion
A. Summary
The understanding of organizational culture related to profit enhancement is a of-
ten discussed but less researched area. The lack in research is to some extend related
to difficulties in approaching the less quantifiable field of culture. Defining culture,
26
measuring it and then linking cultural values to measurement methods of corporate
profitability is still difficult to accomplish. The independent measurement and the
ability of link organizational values to financial data makes this research design espe-
cially interesting. The application of text mining, natural language processing as well
as financial metrics widen both the horizon of current research as well as for the prac-
tical world. By using supervised learning we linked organizational value statements to
corporate finance metrics that describe according to the latest research the financial
performance and profitability of an organization. Given this linkage we were able to
draw an conclusion on which cultural values drive an organization to a more profitable
operation and that the lack of some values in contrast describes organizations with a
lower grade of profitability.
As a result of our analysis we constructed three tree structures that define which
cultural values are on an organizational level the main drivers for higher scores in
profitability measures. The study incorporated three different profitability measure-
ments namely ROE,ROA,E BIT DA/E V . The cultural values that drove most of
the measurements are product related [product, customer, part, ..], incentive related like
[bank, bonus, salary, growth, ...], organizational related [life, clear, decision, balance, ...].
Especially interesting is that beside business related values also values that bridge into
the private life play a role. This are values such as [balance, opportunity, lif e, ...]. The
presented three main groups are more or less representative and important in defining
companies that have high rations in the three different profitability metrics. Especially
the ROE favours companies with a higher willingness to risk and growth. However,
in general shows the list of important organizational cultural values high similarities
between the different measures.
Examine in more detail the sub set of statements that contain the relevant value
stems shows that organizations form the Nordics are in some values over represented
27
such as [goal, employ, balanceorlife] This is well inline with findings from the litera-
ture that described Nordic firms as competitive but also focused on a positive work
life balance as well as valuing personal goals from employees. Furthermore, the anal-
ysis showed that certain sectors are more likely to comply with relevant and positive
associated value stems. Industries that have the need for high skilled labor or very
dynamic industries like the technology sector have most likely values that are positive
related to the defined values. However, it needs to be said at this point that no industry
or company is over represented hence we can stated a low bias when it comes to the
industry or company as a factor for profitability prediction. It is much more important
which cultural values accrue and in which combination.
B. Managerial Implications
Hence management has as one core objective the organizations profitability, it
makes sense to have a closer look at the role of culture. (Kreps, 1990) With this the
study contributes to the practitioners understanding of culture in relation to corporate
profitability. The indication that the analysis gives, enables management to actively
drive their organizations towards higher profitability by shaping an environment which
favors certain relevant values.
Furthermore, the study can support the manager in backing up critical feelings with
a more quantitative approach across multiple European organizations. The hypothesis
that certain cultural values can define more success full (here profitable) companies
was tested in this study and with this can be used to back up managers feeling.
28
C. Future Research
Aiming for a more flexible and adaptive approach in researching organizational
values, this study builds on top of other work from the author in the field of text mining
and leadership, see for this Preuss (2017). The focus of this paper to link culture to
organizational profitability enables future research to look into improvements both in
method as well as the application of this approach onto other fields of organizational
research.
Given the experience from this research, we strongly recommend future research to
look into other critical business objectives and its relation to culture, organizational
values and leadership. Especially situations like restructurings or M&A transactions,
where culture plays according to current research a major role, should be in the focus
of future research in this field. Also differences between industries of regions in which
organizations operate could be an interesting field of research to pursue in the future.
29
VIII. Appendix
A. Contained word stems
The following table presents a small snapshot of the in the under laying data set
contained word stems.
Word Variable Value matrix
activ : num 0 0 0 0 0 0 0 0 0 0 ...
actual : num 0 0 1 0 0 0 0 0 0 0 ...
advanc : num 0 0 0 0 0 0 0 0 0 0 ...
anonym : num 0 0 0 0 0 0 0 0 0 0 ...
averag : num 0 0 0 0 0 0 0 0 0 0 ...
balanc : num 0 0 0 0 0 0 0 0 0 0 ...
base : num 0 0 0 0 0 0 0 0 0 0 ...
benefit : num 0 0 0 1 0 1 1 1 0 0 ...
bonus : num 0 0 0 0 0 0 0 0 0 0 ...
brand : num 0 0 0 0 0 0 0 0 0 0 ...
busi : num 1 0 0 0 0 0 0 0 0 0 ...
care : num 0 0 0 0 0 0 0 1 0 0 ...
career : num 0 0 0 1 0 0 0 0 0 0 ...
challeng : num 1 0 0 0 0 0 1 1 0 0 ...
chanc : num 1 0 0 0 0 0 0 0 0 0 ...
... ... ...
local : num 0 0 0 0 0 1 0 0 0 0 ...
locat : num 0 0 0 0 0 0 0 0 0 0 ...
look : num 0 0 0 0 0 0 0 0 1 1 ...
lot : num 1 0 0 1 0 0 0 0 0 0 ...
lower : num 0 0 0 0 0 0 0 0 0 0 ...
made : num 0 0 0 0 0 1 0 0 0 1 ...
make : num 1 1 0 0 1 0 2 2 2 1 ...
market : num 0 0 0 0 0 0 0 0 0 0 ...
mean : num 0 0 0 0 0 0 0 0 0 0 ...
Table VI: ’data.frame’: 2882 obs. of 177 variables
30
B. Error plots of tree models
The following plots present the relative error rate and the reached R2by number
of splits in each of the tree models. The relative error (RE) graph presents the ratio
between the predicted value and the true value. If both values are the same the outcome
is 1 or 100 percent. The aim is therefore to build a model where the RE fluctuates
around 1. The relative error of the models in figure the shown figures shows a plot
around 1.0. This is for all profitability measures the case.
At this point it is important to mention that in the case of this research, the tree
model is not used to exact predict the performance measure where by a lower accuracy
of the tree model is accepted. It is instead of that only used to select the most important
values that are associated with firm profitability. Since firm profitability is a complex
phenomenon, it can not be expected that the model is able to entirely predict it based
on organizational values.
Figure 4. : Plot tree model ROA
31
Figure 5. : Plot tree model ROE
Figure 6. : Plot tree model EBITDA/EV
32
C. Mathematical concept of model
The used method is based on some mathematical concepts that were influenced by
Quinlan. (Quinlan, 1986) The first step of the process to build the tree consisting of
threshold rules and nodes is named Gini impurity and it can be used to construct a
tree for classification and regression problems. (Mingers, 1989) In this case the aim is
to predict profitability measures which means that we deal with a regression problem
that we want to solve with a tree algorithm. A regression problem means that it solves
the task of approximating a mapping function from input variables (in this case the
word stems and n-grams and their occurrence) to a continuous output variable (in this
case the profitability measures). The Gini impurity measures how often a element from
a data set would be labeled in an incorrect way, if it was randomly labeled by following
the distribution of labels in the data set. It computes a probability piof a data point
with the label i.
X
k6=i
= 1 pi(1)
The equation reaches its minimum when all cases are in the same target category.
The computation of a Gini impurity for Jclasses with i1,2,3, ..., J and piis the
part of items that is labeled with class i
IG(p) =
J
X
i=1
piX
k6=i
pk=
J
X
i=1
pi(1 pi) =
J
X
i=1
(pip2
i) =
J
X
i=1
pi
J
X
i=1
p2
i= 1
J
X
i=1
p2
i(2)
The second part that influence the construction of a decision tree is the calculation
of the information gain. It is used by the latest generations of decision trees.
33
Formula Meaning
IG(T,a) Information gain
H(T) Parent
H(T—a) Weighted sum of children
Table VII: Equal statements in a decision tree
H(T) = IE(P1, P2, ..., pJ) =
J
X
i=1
pilog2pi(3)
In this equation (p1, p2, ...) are fractions that are in it’s sum 1. The fractions
represent the percentage that each class has from the node that results out of the
splits in the tree.
The model will generate the feature accuracy that results in the purest daughter
node. By this we will get a set of rules that describe the accuracy of a features in the
data set in order to get a specific result. In this case of a high or low profitability. In
case of the application on regression problems, also an other factor need to be taken
into consideration and this is the variance.
The reduction of variance in the target variable simply means a sort of discretiza-
tion. The reduced variance in a specific node of the tree (N) is defined as the reduction
of the total variance of the target variable xdue to a split of the current node. It can
be described by the following formula.
Iv(N) = 1
|S|2X
iS
X
jS
1
2(xixj)2(1
|St|2X
iSt
X
jSt
1/2(xixj)2
+1
|Sf|2X
iSf
X
jSf
1/2(xixj)2)
(4)
As with other methods, the input parameters need to be well defined in order to
34
generate meaningful results. In the case of a tree algorithm, one of the parameters that
has to be adjusted are the debt of the tree which describes how many rules the tree
should contain. The tree gets defined by the parameters minsplit: minimal splits of the
tree that should be made to do the classification. The next parameter is minbucket:
minimum number of observations in any terminal node. The last feature that adjust
the design of a tree model is complexity parameter (in the following also named cp)
which plays a role in regression model trees. It says that if the R2is not improved by
the factor that is stated under cp this means that the split is not done.
The above described values that have to be adjusted (e.g. debt of the tree, min-
split, etc.) are named parameters of the model. Those parameters influence the model
accuracy and outcome. The parameters can be set manually but in order to reach the
optimal solution for the underlying problems we incorporate a more scientific approach.
For optimizing the parameter set we use a grid search method. This gird search al-
gorithm was running the unsupervised learning algorithm with different combinations
of parameters to tune the accuracy with which the variables (here word stems) could
define a higher value of the profitability measures. The grid search method builds
for this a grid of the relevant parameters and the different settings of the param-
eters. This can be defined as a nxm matrix with n=numberofparameters and
m=numberof potentialvaluesperfeature. With this the algorithm then selects the
optimal combination of values on each parameter respectively the values on all other
parameters. Given the fact that this method only takes into account the defined pa-
rameters, there is no such optimal solution available, however the above described
approach will lead to a solution which is most close to the theoretical optimal solution.
35
REFERENCES
Aggarwal, Raj, Takato Hiraki, and Ramesh P Rao, 1992, Price/book value ratios and
equity returns on the tokyo stock exchange: Empirical evidence of an anomalous
regularity, Financial Review 27, 589–605.
Bansal, Pratima, 2003, From issues to actions: The importance of individual concerns
and organizational values in responding to natural environmental issues, Organiza-
tion Science 14, 510–527.
Basu, Sanjoy, 1977, Investment performance of common stocks in relation to their price-
earnings ratios: A test of the efficient market hypothesis, The journal of Finance 32,
663–682.
Beaver, William H, and Stephen G Ryan, 2000, Biases and lags in book value and their
effects on the ability of the book-to-market ratio to predict book return on equity,
Journal of accounting research 38, 127–148.
Belgiu, Mariana, and Lucian Dr˘agut¸, 2016, Random forest in remote sensing: A review
of applications and future directions, ISPRS Journal of Photogrammetry and Remote
Sensing 114, 24–31.
Bluedorn, Allen C, Thomas J Kalliath, Michael J Strube, and Gregg D Martin, 1999,
Polychronicity and the inventory of polychronic values (ipv) the development of an
instrument to measure a fundamental dimension of organizational culture, Journal
of managerial psychology 14, 205–231.
Bourne, Humphrey, Mark Jenkins, and Emma Parry, 2017, Mapping espoused organi-
zational values, Journal of Business Ethics 1–16.
36
Cable, Daniel M, and Jeffrey R Edwards, 2004, Complementary and supplementary
fit: a theoretical and empirical integration., Journal of applied psychology 89, 822.
Calori, Roland, and Philippe Sarnin, 1991, Corporate culture and economic perfor-
mance: A french study, Organization studies 12, 049–74.
Carlisle, Tobias E, 2014, Deep Value: Why Activist Investors and Other Contrarians
Battle for Control of Losing Corporations (John Wiley & Sons).
Edmans, Alex, 2011, Does the stock market fully value intangibles? employee satisfac-
tion and equity prices, Journal of Financial economics 101, 621–640.
Fama, Eugene F, 1970, Efficient capital markets: A review of theory and empirical
work, The journal of Finance 25, 383–417.
Fama, Eugene F, 1991, Efficient capital markets: Ii, The journal of finance 46, 1575–
1617.
Gagliardi, Pasquale, 1986, The creation and change of organizational cultures: A con-
ceptual framework, Organization studies 7, 117–134.
Graham, Benjamin, and Bill McGowan, 2005, The intelligent investor (Harper Collins).
Gray, Wesley R, and Tobias E Carlisle, 2012, Quantitative Value,+ Web Site: A Prac-
titioner’s Guide to Automating Intelligent Investment and Eliminating Behavioral
Errors, volume 836 (John Wiley & Sons).
Greenblatt, Joel, 2010, The little book that still beats the market , volume 29 (John
Wiley & Sons).
Grøgaard, Birgitte, and Helene Loe Colman, 2016, Interpretive frames as the orga-
nizations mirror: From espoused values to social integration in mnes, Management
International Review 56, 171–194.
37
Gruman, Jamie A, and Alan M Saks, 2011, Performance management and employee
engagement, Human Resource Management Review 21, 123–136.
Guiso, Luigi, Paola Sapienza, and Luigi Zingales, 2015, The value of corporate culture,
Journal of Financial Economics 117, 60–76.
Gurung, Anil, and Edmund Prater, 2017, A research framework for the impact of
cultural differences on it outsourcing, in Global Sourcing Of Services: Strategies,
Issues And Challenges, 49–82 (World Scientific).
Hofstede, Geert, 1983, The cultural relativity of organizational practices and theories,
Journal of international business studies 14, 75–89.
Hofstede, Geert, Bram Neuijen, Denise Daval Ohayv, and Geert Sanders, 1990, Mea-
suring organizational cultures: A qualitative and quantitative study across twenty
cases, Administrative science quarterly 286–316.
Hothorn, Torsten, Kurt Hornik, and Achim Zeileis, 2006, Unbiased recursive partition-
ing: A conditional inference framework, Journal of Computational and Graphical
statistics 15, 651–674.
House, Robert, Mansour Javidan, Paul Hanges, and Peter Dorfman, 2002, Under-
standing cultures and implicit leadership theories across the globe: an introduction
to project globe, Journal of world business 37, 3–10.
House, Robert J, Paul J Hanges, Mansour Javidan, Peter W Dorfman, and Vipin
Gupta, 2004, Culture, leadership, and organizations: The GLOBE study of 62 soci-
eties (Sage publications).
Howell, Andrea, Andrea Kirk-Brown, and Brian K Cooper, 2012, Does congruence be-
tween espoused and enacted organizational values predict affective commitment in
38
australian organizations?, The International Journal of Human Resource Manage-
ment 23, 731–747.
Janson, Lena, 1994, Towards a dynamic model of post-acquisition cultural integration
.
Kabanoff, Boris, Robert Waldersee, and Marcus Cohen, 1995, Espoused values and
organizational change themes, Academy of Management Journal 38, 1075–1104.
Kreps, Gary L, 1990, Stories as repositories of organizational intelligence: Implica-
tions for organizational development, Annals of the International Communication
Association 13, 191–202.
Kristof, Amy L, 1996, Person-organization fit: An integrative review of its conceptu-
alizations, measurement, and implications, Personnel psychology 49, 1–49.
Lior, Rokach, et al., 2014, Data mining with decision trees: theory and applications,
volume 81 (World scientific).
Liu, Jing, Doron Nissim, and Jacob Thomas, 2002, Equity valuation using multiples,
Journal of Accounting Research 40, 135–172.
Loh, Wei-Yin, 2011, Classification and regression trees, Wiley Interdisciplinary Re-
views: Data Mining and Knowledge Discovery 1, 14–23.
Loughran, Tim, and Jay W Wellman, 2011, New evidence on the relation between the
enterprise multiple and average stock returns, Journal of Financial and Quantitative
Analysis 46, 1629–1650.
Migliore, Laura Ann, 2011, Relation between big five personality traits and hofstede’s
cultural dimensions: Samples from the usa and india, Cross Cultural Management:
An International Journal 18, 38–54.
39
Mingers, John, 1989, An empirical comparison of selection measures for decision-tree
induction, Machine learning 3, 319–342.
Ostroff, Cheri, Yuhyung Shin, and Angelo J Kinicki, 2005, Multiple perspectives of con-
gruence: Relationships between value congruence and employee attitudes, Journal
of Organizational Behavior 26, 591–623.
Piotroski, Joseph D, 2000, Value investing: The use of historical financial statement
information to separate winners from losers, Journal of Accounting Research 1–41.
Preuss, Bjørn, 2017, Text mining and machine learning to capture cultural data, work-
ing paper 2.
Quinlan, J. Ross, 1986, Induction of decision trees, Machine learning 1, 81–106.
Quinn, Robert E, and John Rohrbaugh, 1983, A spatial model of effectiveness crite-
ria: Towards a competing values approach to organizational analysis, Management
science 29, 363–377.
Rokeach, Milton, 1968, Beliefs, attitudes and values: A theory of organization and
change .
Rosenblatt, Valerie, 2011, The impact of institutional processes, social networks, and
culture on diffusion of global work values in multinational organizations, Cross Cul-
tural Management: An International Journal 18, 105–121.
Sarala, Riikka M, and Eero Vaara, 2010, Cultural differences, convergence, and
crossvergence as explanations of knowledge transfer in international acquisitions,
Journal of International Business Studies 41, 1365–1390.
Schein, Edgar H, 2010, Organizational culture and leadership, volume 2 (John Wiley
& Sons).
40
Schwartz, Shalom H, 1992, Universals in the content and structure of values: Theo-
retical advances and empirical tests in 20 countries, Advances in experimental social
psychology 25, 1–65.
Selnes, Fred, Bernard J Jaworski, and Ajay K Kohli, 1996, Market orientation in united
states and scandinavian companies. a cross-cultural study, Scandinavian journal of
management 12, 139–157.
Warner-Søderholm, Gillian, 2012, Culture matters: Norwegian cultural identity within
a scandinavian context, Sage Open 2, 2158244012471350.
Weber, Yaakov, Oded Shenkar, and Adi Raveh, 1996, National and corporate cultural
fit in mergers/acquisitions: An exploratory study, Management science 42, 1215–
1227.
41
... We use the acquisition performance metrics (growth in RoA, growth EBIT and growth CF) as target for each regression model (for each metric one). The interpretation of the results, follows the logic shown in Chapter 3 and 4 (see also Preuss (2018b) and Preuss (2019b)). ...
... The principles underlying the model can be obtained from Chapter 3 and 4, (see also Preuss (2018b) and Preuss (2019b)). More details on the mathematical principles of a regression decision tree can be found in Loh (2014). ...
... According to Gentzkow et al. (2019), nonlinear models are new in economic research. We apply certain steps of sensitivity analysis to build trust in the applied method which are similar to those applied in Chapter 4 (see also Preuss (2018b)). ...
Chapter
With the linkage of acquisition performance metrics to organizational cultural values, the chapter delivers an insight into the role of culture on a firm-specific level. M&A transactions have a high impact on an organization's history and research finds that culture has a great influence on the outcome of these transactions. Hence, this study is valuable. Beside the actual findings, the chapter is an application of a rather new method used in this dissertation for cultural and social science research, called text mining in combinations with machine learning models. Along the analysis, the chapter compares three measures for acquisition performance and its relation to organizational cultural values. The result reveals that elitist values influence value creation after acquisitions. The analysis indicates that an adaptive organizational culture which is opportunity seeking, supports value creation in the M&A context whereas contradicting values can be value-destroying or are at least associated with a lower value creation.
... Organizations which originates from the Nordic countries showed in recent years a strong sustainable performance. Seeing results that show evidence in the linkage between an organizations working culture and a company's financial profitability Preuss (2018b), leads us to the question whether there are certain cultural values that differentiate Nordic organizations from other European companies. The opportunities to research values by utilizing modern text mining and machine learning technology makes this research also from a method perspective valuable. ...
... (Hofstede, Neuijen, Ohayv, and Sanders, 1990) A number of recent studies from the author have examined the potential of these new methods in connection to cultural research. (Preuss, 2017b) (Preuss, 2017a) (Preuss, 2018c) (Preuss, 2018a) (Preuss, 2018b) Having said this the latest approaches use less questionnaires to collect data and then build frameworks on cultural dimensions with all the related critics. (Schwartz, 1992a) In contrast they use a data driven approach and instead of collecting information specifically for the purpose of research, already existing information is used. ...
... Before using the text in quantitative models, we have to transform the text into a for a computer readable format. We do this by applying certain techniques around natural language processing and text mining, which were previously used and described in Preuss (2017b) and Preuss (2018b). At the end of the whole processing steps we have a table or matrix of words and word combinations in a homogeneous form which can be interpreted by a statistical model. ...
Preprint
Full-text available
With the goal of this paper in analyzing the cultural differences between organizations in the Nordics and Europe, we aim to contribute to the research on working culture and in particular the working culture in Scandinavian organizations. Throughout the utilization of machine learning and text mining, we analyze a large corpus of textual data and measure which of the inherited organizational values differentiate Nordic organizations form other European organizations. In the results we found evidence that personal values and long-term orientation is more significant for organizations of the Nordics then career and profit orientation. We support with these findings current research and establish a new way to analyze qualitative information on an organizational level.
Chapter
Full-text available
This thesis presents a set of seven chapters that can be grouped into two sections namely Part I-what can a new measurement of culture look like? and Part II-which roles do culture play in a firms economy? The chapters cover the description of organizational values, the impact of organizational culture on a company's profitability, the role of culture in acquisitions or the country-specific differences in corporate culture. All chapters except for Chapter 6 use a new method for quantifying textual data. They incorporate natural language processing, techniques from text mining, non-linear modeling and machine learning. With this, we can summarize the chapters under the title NLP as a way to analyze corporate culture.
Chapter
Full-text available
The increasing amount of data and the importance of soft factors affecting economic prosperity make this work an interesting and important read. The chapters in this book have in common that natural language processing helps to understand organizational cultural values. The first part discusses and forms a new method to analyze cultural values derived from textual statements. Here, we discuss the applicability of natural language processing (NLP) in economic research. The second part applies the method of NLP in organizational cultural values and links the measurement of culture to economic metrics from profitability to M&A success measurement. The results show the importance of product-, organizational- and incentive related values to form a favorable environment for economic prosperity.
Chapter
In this chapter, we investigate whether culture-related difficulties arise in the merger between Maersk and Hamburg Sued. Culture has a major impact on the success of M&A transactions and can make or break a merger or acquisition. In this chapter we analyze which cultural values might play a major role in this case. The difference of organizational culture might further more influence cross-border merges however, specific culture values might be relevant too. The application of Hofstede's dimensions of organizational culture makes the results comparable with other scholars' findings. Setting out to find which cultural elements influence the acquisition of Hamburg Sued through Maersk, we describe both parities organizational culture and highlight critical points. We furthermore reflect on the influence that these critical points have on the acquisition. The primary factor remains the flexibility of a culture in order to complete a successful acculturation process.
Chapter
The following chapter develops a linkage between organizational cultural values and firm profitability. After reviewing the literature, it becomes clear that culture plays a minor side role in the analysis of firm profitability. However, scholars mention the impact of cultural values on organizational development which is a core motivation for this study. We build on literature that state a strong linkage between firm performance and organizational culture. We base the construction of profit-enhancing organizational values on a data set of ca. 2.800 value statements of ca. 150 organizations in Europe. We link the values to three different profitability measures that are derived from current research. We analyze which of the values have the strongest association to profitability and with that characterize a profitable firms organizational culture. The results show the domination of business values that are customer-focused and product-related. In addition, values related to human resource play a significant role. By advancing our understanding of organizational culture and its linkage to a firms profitability, we provide a means by which future research into organizational values and corporate culture can progress.
Article
Full-text available
This paper develops an inventory and conceptual map of espoused organizational values. We suggest that espoused values are fundamentally different to other value forms as they are collective value statements that need to coexist as a basis for organizational activity and performance. The inventory is built from an analysis of 3112 value items espoused by 554 organizations in the UK and USA in both profit and not-for-profit sectors. We distil these value items into 85 espoused value labels, and these are assessed in terms of their similarity and difference through judgements made by 53 experienced individuals. The resulting conceptual map facilitates the evaluation of values which are espoused at the organizational level, as opposed to aggregations of personal values, an important distinction that is often ignored in the literature. This analysis identifies a number of distinct areas of emphasis occupied by espoused values. In particular, the richness of value labels that relates to broader ethical issues may be aimed at external stakeholder management, but also may have an increasing influence on organizational behaviour as they are embedded into organizational practices. By advancing our understanding of espoused values, through an analysis of those being used in practice, we provide a means by which future research into organizational values and ethical issues can progress.
Article
Full-text available
Many multinational enterprises (MNEs) seek to strengthen their competitive positions through internal integration. Socialization is a key integration mechanism to leverage advantages across MNEs’ geographically dispersed organizational units. Parent organizations often communicate a set of values intended to guide action throughout the MNE, referred to as espoused values, to initiate a socialization process. However, we have limited insights into how espoused values are endorsed and subsequently contribute to MNE integration. Through a case study, we analyze how espoused values are interpreted by the foreign subsidiaries and influence subsequent subsidiary behavior. Our findings suggest that the socialization process is complex, where the local context and perceptions of headquarter nationality provide the frames for interpretation. We identify that even though the espoused values may differ in their operationalization in local contexts, they can still contribute to MNE integration. This study contributes to existing MNE literature by conceptualizing the role of interpretive frames and the endorsement of values in achieving integration through espoused values.
Article
Full-text available
Contracting Global Virtual Teams as part of global IT outsourcing is currently en vogue. As might be expected when virtual team members are from different countries, cultural factors play an important role in the success of outsourcing. However, there have been very few studies that assess the effect of culture on IT outsourcing and virtual teams. This conceptual paper addresses this oversight by looking at the effect of cultural differences on IT outsourcing and virtual teams' performance. The applicable literature on outsourcing, virtual teams and culture is analyzed and a framework of offshore outsourcing success is developed. This framework includes the concept of psychic distance to better understand the phenomenon of virtual teams and outsourcing success. Adding this as a key research component provides a more realistic way of researching global virtual teams. Future directions for research based on the developed framework are also provided. By assessing the cultural differences of virtual teams in IT outsourcing, our research framework will help academics pursue this growing business phenomenon.
Article
Full-text available
Clustering algorithms have emerged as an alternative powerful meta-learning tool to accurately analyze the massive volume of data generated by modern applications. In particular, their main goal is to categorize data into clusters such that objects are grouped in the same cluster when they are similar according to specific metrics. There is a vast body of knowledge in the area of clustering and there has been attempts to analyze and categorize them for a larger number of applications. However, one of the major issues in using clustering algorithms for big data that causes confusion amongst practitioners is the lack of consensus in the definition of their properties as well as a lack of formal categorization. With the intention of alleviating these problems, this paper introduces concepts and algorithms related to clustering, a concise survey of existing (clustering) algorithms as well as providing a comparison, both from a theoretical and an empirical perspective. From a theoretical perspective, we developed a categorizing framework based on the main properties pointed out in previous studies. Empirically, we conducted extensive experiments where we compared the most representative algorithm from each of the categories using a large number of real (big) data sets. The effectiveness of the candidate clustering algorithms is measured through a number of internal and external validity metrics, stability, runtime, and scalability tests. In addition, we highlighted the set of clustering algorithms that are the best performing for big data.
Article
Full-text available
Fifty years ago there were just a handful of universities across the globe that could provide for specialized educational courses. Today Universities are generating not only graduates but also massive amounts of data from their systems. So the question that arises is how can a higher educational institution harness the power of this didactic data for its strategic use? This review paper will serve to answer this question. To build an Information system that can learn from the data is a difficult task but it has been achieved successfully by using various data mining approaches like clustering, classification, prediction algorithms etc. However the use of these algorithms with educational dataset is quite low. This review paper focuses to consolidate the different types of clustering algorithms as applied in Educational Data Mining context. Index Terms—Clustering, educational data mining (EDM), learning styles, learning management systems (LMS).
Article
A random forest (RF) classifier is an ensemble classifier that produces multiple decision trees, using a randomly selected subset of training samples and variables. This classifier has become popular within the remote sensing community due to the accuracy of its classifications. The overall objective of this work was to review the utilization of RF classifier in remote sensing. This review has revealed that RF classifier can successfully handle high data dimensionality and multicolinearity, being both fast and insensitive to overfitting. It is, however, sensitive to the sampling design. The variable importance (VI) measurement provided by the RF classifier has been extensively exploited in different scenarios, for example to reduce the number of dimensions of hyperspectral data, to identify the most relevant multisource remote sensing and geographic data, and to select the most suitable season to classify particular target classes. Further investigations are required into less commonly exploited uses of this classifier, such as for sample proximity analysis to detect and remove outliers in the training samples.