revenue diversification in dutch charity …
TRANSCRIPT
REVENUE DIVERSIFICATION IN
DUTCH CHARITY ORGANIZATIONS:
DOES IT LEAD TO GROWTH?
by Femke de Jong
s856006
Master Thesis
Tilburg University
School of Economics and Management
Finance Department
Supervisor: dr. D.A. Hollanders
2
Revenue Diversification in Dutch Charity Organizations: Does it lead
to Growth?
Name: F. (Femke) de Jong
Student number: s856006
Supervisor: dr. D.A. Hollanders
Date of Defense: October 7th, 2014
Session Chair: Dr. A. Manconi
Faculty: Tilburg School of Economics and Management
3
Abstract
This master thesis investigates the relationship between revenue diversification and
growth. To investigate this, a dataset composed out of data provided by the Central Bureau of
Fundraising (CBF) is used. This dataset consists of financial data of 1,282 Dutch charity
organizations. A panel data model with fixed effects is used to test the hypothesis whether there
is a relation between diversification and growth. The results suggest that revenue diversification
turns out to have a negative impact on growth, even when controlled for several other factors
which might influence this. Reason for this negative relationship, in contrast to a positive
relationship that other scholars find, might be the differences between the for- and the non-profit
world.
Preface
I would like to thank David Hollanders for supervising me during the process of writing
this master thesis, his critical notes and useful help. I would also like to thank Jos Grazell, for
helping me at the start of this process and discussing the subject with me. Furthermore, I would
like to thank Ad Graaman and Fred de Jong of the CBF, for providing me all the data I needed.
Next to that, I would like to thank my fellow students and my friends, who made my
student life incredibly wonderful, in every kind of way. I will definitely miss it. I want to thank
Henk, for all his support and critical questions. And last but not least, I would like to thank my
parents, Anne & Jikky, who made it possible for me to study, to develop myself during these last
six years and gave me the opportunity to get the most out of this period.
4
Table of Contents
Introduction ..................................................................................................................................... 5
I. Literature review ..................................................................................................................... 7
i. Diversification..................................................................................................................... 7
Corporate diversification ........................................................................................................ 7
Financial diversification........................................................................................................ 11
ii. Diversification in relation to Non-profit firms and Growth.............................................. 13
Diversification in the non-profit sector ................................................................................. 13
Diversification and growth ................................................................................................... 16
II. Research Methodology ......................................................................................................... 17
III. Data ....................................................................................................................................... 20
IV. Descriptive statistics ............................................................................................................. 22
V. Results ................................................................................................................................... 25
VI. Conclusions ........................................................................................................................... 32
VII. Discussion and recommendations ......................................................................................... 34
REFERENCES ............................................................................................................................. 36
APPENDICES .............................................................................................................................. 38
5
Introduction
The financial crisis that hit the world in 2008 had not only impact on the world economy,
but also on the economic climate of the Netherlands. As a result, among many other
consequences, the number of bankruptcies increased, unemployment rates grew and government
expenditures were cut. In 2013, the real disposable income of the average Dutch household
decreased for the sixth year in a row (Centraal Bureau voor de Statistiek [CBS], 2014). Among
the first expenses, households intend to cut their donations to charities. This does not only occur
on micro level at households, also governments cut their expenses on development aid at large
scale; from 0.8 percent of the Gross National Product (GNP) in 2008 to 0.7 in 2013
(Rijksoverheid, 2014). As the GNP has been decreasing since the start of the economic crisis as
well, the budget for development aid has decreased even more. These economic and social trends
cause a more competitive and complex environment for funding of charity projects. Therefore, a
sound financial base is more and more important to charity organizations. One of the broadly
accepted theories about how non-profits can manage their financials in a sustainable way is the
concept of revenue diversification (Frumkin and Keating, 2011).
The Dutch charity organizations do not only rely on donations from individuals and
government subsidies. The Central Bureau of Fundraising (CBF), a Dutch independent
foundation that collects and provides information about fundraising organizations in the
Netherlands, identifies 10 different revenue streams, which include – next to subsidies and
donations – for example mailings, legacies and investment income (CBF, 2014). Spreading your
revenue income on several different revenue streams is called revenue diversification. When one
does not diversify, one bears the risk to have zero income when that particular source of income
disappears. By diversifying, one can spread this risk of running out of income. This concept of
diversification is an old concept that is not only applicable for revenue sources or for the non-
profit world. The concept is widely applied within the financial world, where diversifying can be
defined as a management technique where one chooses a wide range of different investments
within a portfolio. By doing this, such portfolio will, on average, yield higher returns and pose a
lower risk than any individual investment found within the portfolio (Bodie, Kane, & Marcus,
2011). Many scholars investigated this concept and its implications; the consequences for costs
and benefits of the firm, the stability of the portfolio and shareholder and firm value.
6
Diversifying is also an applied concept at corporate firms. Many firms choose to diversify
in terms of merging and acquiring, going international or enter different product markets
(corporate diversification). Since the 90s, the concept of diversification is also applied in the
non-profit sector. Scholars used modern financial concepts, like diversification, to model the
financial situation of non-profit organizations. One needs to realize the duality of non-profit
organizations when applying these ‘general’ financial models to the non-profit world: “Non-
profit organizations often face the dual task of achieving mission-related goals while maintaining
a healthy financial condition that ensures organizational survival.” (Carroll and Stater, 2008,
p. 947).
Much research has been executed to review the relation between revenue diversification
and the stability of the firm. It is concluded that by spreading their risk among multiple revenue
sources, a non-profit reduces its revenue volatility and strengthens its financial position. Whether
such multiple revenue sources lead also to an increase of the total revenue, is an underestimated
subject. When an organization employs an additional revenue source, and this revenue source
brings a lot of additional income, the total revenue will grow and thereby the organization may
also grow. From that perspective, diversification may lead to growth.
Whether it is ethically right to pursue growth as a charity organization is a justified
question. Therefore, a distinction has to be made between growing of the charity organization
itself (the overhead) and the mission fulfillment of the organization. Charity organizations that
grow (only) in overhead costs should look into ethical aspects of their future concerning their
mission and social responsibility. Nevertheless, to fulfill its mission, a charity organization will
have to raise enough funds. Growth in revenue is therefor, in principle, not a negative thing, but
one needs to spend their funds wisely in respect to a sustainable future position.
To accomplish sustainable growth, one cannot simply utilize more income sources and
therefor gain more income. Managing more different sources involves, for example, fundraising
and administrative costs, which may exceed the benefits. The relation is perhaps not as easy as it
sounds. Therefore, this thesis focusses on the relation between revenue diversification and
growth and investigates whether non-profit organizations that diversify more experience more
7
growth. Hence, the main question which will be answered in this master thesis is: Revenue
Diversification in Dutch Charity Organizations: Does is lead to Growth?
This master thesis is organized as follows: The thesis first contains a brief exploration of
the several disciplinary perspectives of diversification in the current literature (section I).
Afterwards, the research methodology (section II) and the data are described (section III), the
descriptives and results are given (section IV and section V). Finally, the results are compared to
the existing literature and the final conclusion is provided (section VI) and discussed with
recommendations for further research (section VII).
I. Literature review
Diversification is a well investigated subject in the economic literature. Much research
has been done about the motives to diversify and the impact it has on the firm. Diversification
knows many forms in many different fields. In this review, first the main concepts about
diversification in these different fields will be explained. Secondly, the diversification concept
will be applied to the non-profit world and linked to growth.
i . D IVERSIFICATIO N
The concept of diversification can be split into two different broad concepts: Corporate
and Financial diversification. Corporate diversification is defined by Ramanujam and
Varadarajan (1989) as follows: “The entry of a firm of business unit into new lines of activity,
either by processes of internal business development or acquisition, which entail changes in its
administrative structure, systems, and other management processes.” (p. 525). Financial
diversification, on the other hand, is defined as a risk management technique that mixes a wide
variety of assets in a portfolio, so that the exposure to the risk of any particular asset is limited
(Bodie et al, 2011). Therefore, a certain company can choose to diversify on both corporate and
financial level. Both concepts and its implications will be explained and analyzed in this section.
CORPORATE DIVERSIFICATION
Corporate diversification and its effects on a firm’s value is a long-standing controversy
in the literature (de Andrés, de la Fuente, & Valesco, 2014). On the one hand, a large body of
literature states that there is a so-called diversification discount, but on the other hand, some
8
scholars suggest a diversification premium. Based on the literature investigation done by Martin
& Sayrak (2003), this contradiction will be studied and explained.
To start, diversification brings, naturally, costs and benefits. One potential cost arising
from diversification results from the agency theory: managers choose to diversify out of self-
interest, ignoring the interests of the stockholders. They expect the pure pleasure of empire
building; namely, that it will increase their compensation (Jensen and Murphy, 1990), power and
prestige (Jensen, 1986). Next to that, managers want to make themselves indispensable by
making investments that require their particular skills (Shleifer and Vishny, 1989). This
self-enriching choices are not optimal for an organization and therefor a cost. Thereby, managers
tend to over invest, especially when they have access to an internal capital market. When an
internal capital market is created, a firm’s cash flows generated by one segment can be used to
finance another segment (cross-subsidization) and, likewise, a segment’s assets can be used as
collateral for obtaining funding for other segments (Erdorf, Hartmann-Wendels, Heinrichs, &
Matz, 2013). Cross-subsidization is a positive thing in itself, but when a firm has excess or free
cash flow, it gives a greater opportunity to over invest (Martin and Sayrak, 2003). Next to these
agency problems, an inefficiency problem is recognized. Diversified firms may not have more
free cash flow, but simply do a worse job of allocating their resources than focused firms. This
problem can also be identified as a cost of cross-subsidization (Erdorf et al., 2013). It is possible
that this problem arises due to information asymmetry problems between the firm’s central
management and the management of the operating divisions (Harris, Kriebel & Raviv, 1982).
When the agency theory and the inefficiency problem are linked, it might imply that managers
decide to transfer and invest the money from internal capital markets in a value destroying way.
As mentioned, an internal capital market is a positive thing and can, next to the costs,
bring benefits to the organization. An internal capital market offers a number of possible sources
of value to the firm’s owners (Martin and Sayrak, 2003): internally raised equity capital is less
costly than funds raised in the external capital market, due to transaction and information costs.
Thereby it gives managers decision control instead of the investors. Managers can do a better job
of project selection (“winner picking”) and thus enhance film value (Stein, 1997). Having
internal capital markets gives also the opportunity to transfer money from operating divisions
with limited opportunities to others that are more promising to create shareholder value. Thereby,
9
cross-subsidization can be efficient if it helps the firms eliminate some of the costs of financial
constraints (Erdorf et al, 2013). Yet another benefit of diversification is formed by the possibility
to reduce the variance of future cash flows by diversifying activities at firm level. This
diversification serves to increase the firm’s debt capacity, since creditors may rely on the
combined fortunes of all the diversified firm’s units, which adds value to the firm (Lewellen,
1971).
But to what extent do these costs and benefits create or destroy shareholder value? When
diversification destroys shareholder value, the shares of a diversified firm sell at a discount;
when diversification creates shareholder value, the shares sell at a premium. Over the 20th
century, the existence of a diversification discount or premium is examined a lot. Martin and
Sayrak (2003) review the literature and structure this in three “rounds” of waves of research.
The first round forms the basis for the present “consensus” among most financial
economists: corporate diversification destroys value. The evidence that supports this conclusion
comes from a variety of sources: to start, diversified firms tend to have a lower Tobin’s Q1 then
specialized firms do (Lang and Stulz, 1994). This difference occurs after controlling for firm size,
research & development and access to financial markets (Lang & Stulz, 1994). Secondly,
diversified firms trade at discounts of up to 15% when compared to the value of a portfolio of
comparable stand-alone firms. This effect is even larger for firms who diversify in unrelated
markets. Inefficient cross-subsidization might account for a part of this discount (Berger & Ofek,
1995). Third, diversified firms face an increased likelihood of being broken up through
reorganization (Berger & Ofek, 1996) and finally, firms who increase their focus by selling part
of their assets (which can be seen as the opposite of diversification), experience an increase in
stock price returns after the announcement (John & Ofek, 1995). The reasons of these mentioned
value destructions are divided in two possibilities and in line with the costs of diversification
stated before: either diversified firms take inefficient action in their allocation of internal
generated funds, or poor allocation are made on purpose due to agency problems. In both cases,
these problems result in non-optimal cross-subsidization, where the firm’s weaker divisions are
supported with investments from cash flows from stronger divisions (Martin and Sayrak, 2003).
1 Tobin’s Q is defined as the market value of the firm divided by the book value of assets (Lang and Stulz, 1994)
10
In the second round, a number of studies begun to argue that the discount is attributable
to factors other than diversification. The notion that diversified firms sell at a discount is not
contested, but it is argued that the discount does not arise due to diversification, but is a result of
the acquired or acquiring firm selling at a discount prior to merging (Martin and Sayrak, 2003).
They doubt the endogeneity of the diversification decision: the discount is caused by the
systematic difference between the firms that choose to diversify and the typical focused firms
and the diversified firms tend to trade at a discount prior to diversifying (among others, Graham,
Lemmon, & Wolk, 1999; Chevalier, 2000). Next to that, when one controls for fundamental
differences between conglomerate firms and single segment firms in terms of size, capital
expenditures/sales, EBIT/sales, industry growth rate and R&D/sales, the diversification discount
drops or disappears entirely (Campa and Kedia, 2002).
During the third and last round it is argued that there is no diversification discount, and in
fact, diversified firms trade at a significant premium. Differences compared to these and previous
results are attributed to the possibility of measurement errors in prior research (Martin & Sayrak,
2003). Villalonga (2000, 2004) retests the diversification discount hypothesis by correcting for
thee data limitations, since it is believed that the previous attempts to assess the diversification
discount are flawed by these limitations: (1) the extent of disaggregation in segment financial
reporting is less than the true extent of firm diversification such that firms are actually more
diversified than is indicated in segment financial reporting. Thereby, (2) the definition of a
business segment is so flexible that it allows firms to combine two or more activities that are
vertically related into a single segment. Finally, (3), some industries are fundamentally
composed of segments of diversified firms (Villalonga, 2000). When controlling for these
problems, diversified firms trade at a significant premium, not at a discount, when compared to
non-diversified firms from the same industry (Villalonga, 2000, 2004). Reason for this premium
is formed by the argument that that diversified firms may have better access to capital markets
than focused firms, due to valuation problems faced by investors in the presence of asymmetric
information (Hadlock, Ryngaert and Thomas, 2001)
All the studies described above, show a lot of conflicting results and interpretations. This
can be explained by the fact that these studies estimate the effect of diversification on
performance on a large variety of industries (Santalo and Beccera, 2008). These studies do not
11
take into account the possibility that the diversification could confer a competitive advantage on
some industries but not in others. Since they do all measure the average effect of diversification
on performance, this could lead to different conclusions. Santalo and Becerra (2008) show that
the effect of diversification on performance is not heterogeneous across industries. They report
clear evidence that diversified firms observe a diversification discount if, and only if, they
compete in industries where focused firms hold a considerable market share. The rationale
behind this is that the focused firms will have a competitive advantage over diversified firms
driven by a process affiliated to natural selection. There are two reasons for this heterogeneity:
first, in industries in which soft information is important, diversified firms might gain a better
financial performance because of better excess to financial resources. As reason they point out
that the soft information of a company might be easier to transmit inside rather than outside of
the company, and therefore, corporate headquarters of a conglomerate could have access to
valuable information unavailable to external capital markets. Second, they state that the market
structure of vertically connected industries will influence the diversification advantage.
Vertically integrated firms might enjoy a larger competitive advantage over specialized
companies in more concentrated industries because of their lower transaction costs in dealing
with industries with only a few players (Santalo and Beccera, 2008). In their paper, Santalo and
Becerra (2008) just show that the competitive advantage is not homogenous across industries;
they do not identify the key industry characteristics that determine the competitive advantage of
diversification. These characteristics are still not identified present-day and are still an interesting
topic for future research (Erdorf et al., 2013).
F INANCIAL DIVERSIFICATION
The first formal portfolio selection model including diversification is from 1952
(Markowitz). This model describes the first step of the process of investment portfolio selection:
the efficient frontier of risky assets. This frontier is based on the traditional risk-return
relationship combined with the law of large numbers and shows the portfolio with the highest
possible return given a certain level of risk, or, alternatively, the lowest possible level of risk (the
variance) given a certain return (Bodie, Kane, & Marcus, 2011). By keeping a diversified
portfolio with many assets with a lowest mutual correlation, one can derive a portfolio that keeps
actual returns close to the amount of anticipated returns (Markowitz, 1952). To derive this
12
optimal portfolio, one has to keep in mind that not only the characteristics of the individual
securities have to be considered, but that an investor should consider how each security
co-moved with all other securities (Markowitz, 1952).
The most striking conclusion after Markowitz’ work is that each rational portfolio
manager will face the same optimal risky portfolio, since this portfolio maximizes the
reward-to-risk ratio, which is represented by the Sharpe ratio (Sharpe, 1966). This ratio is for
each (rational) portfolio manager the same, since the risk aversion level does not matter in this
model (Bodie, Kane, & Marcus, 2011). Tobin (1958) came up with the separation theory, which
describes the two-step procedure of the portfolio choice problem: the first one leads indeed to the
same portfolio, but in the second step, the portfolio manager has to choose the optimal
combination between risky assets and a risk free investment, which depends on his risk aversion
level. This level is represented by the utility function (indifference curve) and causes different
optimal portfolios for each different level of risk aversion. Next to considering the risk and return
of a portfolio, later research provided other measurements to describe the distribution of the
portfolio; like the systematic risk (Friend and Blume, 1970) and skewness (Lee, 1977).
The theory above describes a solution for a single-period period, but what if the true
problem an investor faces is a multi-period problem? Many scholars analyzed this problem and
found that the multi-period problem can be solved as a sequence of single-period problems,
under several sets of reasonable assumptions (Elton & Gruber, 1997). The optimal portfolio in
the multiple-period case will be different from the single-period case, since the utility function
changes when the investment period changes. Another problem which arises with investigating
multi-period investing is the question whether the returns and variances are correlated.
To derive the optimal portfolio, one needs to estimate the financial data, like covariances,
which is used as input for the model. Estimating this data was an enormous work before the
development of factor models. These models explain the relationship between the rate of return
of a security with a certain factor. The earliest factor model was the single index model, which
was developed and popularized by Sharpe (1967). This model regresses the expected return of a
security on the excess return of a market index. The slope of this regression is the Beta, which
measures the security’s sensitivity to the index. This sensitivity describes the systematic risk; the
13
risk inherent to the entire market or an entire market segment, which cannot be diversified away.
The other part of the total risk a security occurs, non-systematic or non-market risk, can be
diversified away if one increases the number of securities in his portfolio.
The financial diversification theory is a difficult theory with many implications. But
understanding the importance of diversification, can help financial managers to achieve superior
performing portfolio’s (Sorensen et al., 2004). Not only financial managers in the for-profit
world, but also financial managers of non-profit organizations may benefit from applying these
theories wisely.
ii . D IVERSIFICATION IN RE LATION TO NON-PROFIT FIRMS AND GROWTH
In this section, the features and characteristics of both corporate and financial
diversification will be applied to the non-profit world and linked to growth.
DIVERSIFICATION IN THE NON-PROFIT SECTOR
As described above, many economists throughout the years, have modeled how one can
derive an optimal investment portfolio and thereby maximize return while minimizing financial
risk. These methods are particularly applicable for for-profit organizations. To bridge this gap
between the traditional portfolio theories and the non-profit world, one has to take in mind the
differences and similarities between non-profit and for-profit managers (Kingma, 1993). Just like
for-profit managers, non-profit managers choose optimal combinations of risk a return by
selecting different streams of financing. In the non-profit world, managers want to provide a
certain level of services, comparable to a certain level of return, while minimize unpredictable
changes in these services (risk). The risks a non-profit manager bears, are different for the
different revenue streams (Kingma, 1993); for example, for donations, a non-profit organization
risks fundraising expenses; for subsidies, a non-profit organization risks the expenses which one
will occur when complying with the government standards. Besides these risks, the non-profit
manager may run an additional risk in terms of future funding: a non-profit manager may, by
requesting funding from a particular source, not only risk the expense of a promising service but
also the risk of not receiving the additional expected funding at all (Kingma, 1993). Opposed to
for-profit, non-profit organizations differ in their capacity to absorb unexpected changes in
revenues (Kingma, 1993); both in delay of certain revenue streams, like government subsidies
14
(Grossman, 1992), and also the finding that the revenue streams may be lower than expected. But
notwithstanding this last difference, the decision that the non-profit manager needs to make on
his combination of revenue sources, can be seen as a typical risk-return problem, coming from
the risk-return trade-off, which is defined by Kingma (1993) as: “Any increase in expected
revenues, whether from an increase in the time and effort devoted to fundraising, grant
applications, or requesting additional government revenues, is subject to an increase in risk” (p.
109).
Although one has to be careful about applying traditional portfolio theory to the
non-profit setting, the complications are conquerable (Jegers, 1997). The managerial preferences
that Tobin (1958) suggested by the separation theory (introduced earlier in this section), can also
be included in Kingma’s portfolio model. (Jegers, 1997). By introducing this concept, Jegers
(1997) differs from Kingma (1993), since the latter assumes that the expected returns of the
different revenue streams are equal. With this equality assumption, managerial risk aversion
becomes irrelevant (Jegers, 1997). Therefore, the equality assumption is left behind, and the
interaction between managerial risk preferences and the optimal risk-return combination is
included (Jegers, 1997). Next to this assumption, some other differences between the non-profit
world and traditional portfolio theory have to be considered (Jegers, 1997). First, in contrast to a
typical investor who is allowed to lend or borrow with a certain risk-return relationship, the
revenue (or service level) of a non-profit manager is uncertain. Secondly, an investor has only
uncertainty about the return that he receives on his investment, but the non-profit manager knows
uncertainty about the whole revenue. Third, the investor can choose his optimal portfolio; he can
determine which part of his investment he wants to invest in the particular assets. The non-profit
manager, in contrast, can only hope that, the share that he expects from a certain source is equal
to the optimal combination that he desires. Finally, the non-profit manager does not face an
unconstrained optimization problem as the investor in the traditional finance theory does, since
he may be restricted for legal or financial reasons (Jegers, 1997).
The financial concept of diversification appears to be very applicable to the non-profit
world. Also the concepts of corporate diversification can be applied to the non-profit situation.
Just as for-profit firms, non-profit firms may have to deal with self-enriching managers:
managers who do not have the mission of the organization in mind, but differ in terms of empire
15
building. The typical agency problems can also play a role for non-profit organizations. The most
non-profit organizations will also have an internal capital market: a part of the revenues retrieved
will be unrestricted and can be used wherever in the organization (this in contrast to
(semi-) restricted funds which restrict the organization to inject the funds in to several projects of
parts of the organization). As for for-profit organizations, cross-subsidization brings several costs
and benefits to the non-profit organization: the inefficiency problem may play a role here;
comparing the services of a non-profit organization with the return of a for-profit organization, a
firm can allocate the funds to a project which may not give the highest return in terms of
achieving the mission of the organization. But a large internal capital market gives the
organization chances to fulfil several projects.
As described before, the effects of diversification among for-profit organizations are
investigated widely. Carroll & Stater (2009) examine the effects of revenue diversification
among non-profit organizations, and in particular, if diversification leads to greater revenue
stability over time. Stable, healthy non-profit organizations will be more capable of continuing to
work toward their missions and financial stability over time will lead to greater ability to provide
programs, compensate staff and promote mission awareness (Carroll & Stater, 2009). When
measuring the impact of diversification on revenue volatility (or equally, revenue stability), it is
suggested that organizations with a higher diversified revenue portfolio, experience lower
revenue volatility, which implies a stable organization. Revenue diversification, or in other
words, equalizing the reliance on the several different revenue streams, is a viable strategy for a
stable organization (Carroll and Stater, 2009). Carroll and Stater (2008) estimate also the effect
of organizational efficiency and financial flexibility on revenue volatility. About the effect of
organizational efficiency exist several conflicting beliefs: some argue that having lower
efficiency, measured by relative higher non-programmatic expenses to total expenses, leads to
less trust and therefor discourage gifts. Others argue that having less non-programmatic costs
reduces organizational capacity (a.o. Bowman, 2006) and experience fewer program and funding
disruptions (Keating et al. 2005). Carroll and Stater (2008) take over the first argument:
“Organizations that have less administration and fundraising costs, are able to spend more
resources into mission fulfillment, which increases their perceived effectiveness and
consequently their income potential” (p. 954). Concerning financial flexibility, Carroll and
Stater (2008) state that an organization having greater financial flexibility, detectable by greater
16
equity balances and higher operating margins (Chang and Tuckman, 1994), has better
opportunities to engage in future financial planning and to reduce uncertainty during the annual
budget process.
DIVERSIFICATION AND GROWTH
Although there has been a wide spread research about diversification and its impact on
firm value, the relationship between diversification and a firm’s growth opportunities is a less
exposed subject (Andrés et. al, 2014). The growth opportunities of a firm are one of the factors
that may influence the effect of diversification on firm value, and may cause the conflicting
opinions about diversification creating value or not, like the factor as industry named before. The
factor growth opportunities in relation to diversification is yet underexposed. Some scholars
investigated this subject (Andrés et al., 2014) but their contributions are contradictory. Bernando
and Chowdhry (2002) explain the diversification discount by stating that single segment firms
have a lot of growth opportunities, but firms who already diversified, already exploit some of
these. Ferris, Sen, Lim and Yeo (2002) state that diversification destroys value for firms with a
weak cash flow position and low growth opportunities available. In contrast, Stowe and Xing
(2006) show that the diversification discount still remains after controlling for growth
opportunities. Andrés et al. (2014) then show that the effect of diversification on firm value is
contingent on growth opportunities; growth opportunities has a mediating effect. Their results
provide evidence for a quadratic relationship between diversification and growth. This
relationship can be interpreted as that when a firm chooses to diversify in early stages, this
involves replacing growth opportunities by assets in place. However, in later stages,
diversification becomes a net source of further growth options (Andrés et al., 2014).
Just like an underexposed relationship between diversification and growth in the
literature in the for-profit world, there is also done little research about the growth of the non-
profit firm. Scholars investigate factors that influence the growth of the non-profit sector, like
Corbin (1999), but there is less to find specific for a non-profit firm. Frumkin & Keating (2011)
investigate the risks and rewards that revenue concentration (as the opposite of diversification)
can bring. They state that at the time their paper was published, one did not explore weather
revenue concentration had any positive effects, or in other words, what the positive effects of
risk-seeking behavior are in terms of growth and efficiency. By diversifying revenue streams
17
across many sources of funding, non-profits may let slip some opportunities that come from
capitalizing on a particular segment or funding market, which may result in growth or limited
overhead costs, which arise from multiple funding streams (Frumkin & Keating, 2011). They
derive this link with growth from the thought that by concentrating, non-profits are in a position
to develop specialized skills that will enable these managers to be more effective at fundraising
or obtaining government contracts, which will lead to the development of a reputation and long-
term marketing relationships. The link with efficiency is derived from the thought that revenue
concentrators may experience lower overhead and administrative costs, since they need to
manage less revenue streams and there accompanying (complex) contracts and contributions.
After testing the implications of revenue diversification on efficiency gains (i.e. relative limited
costs) they confirm their hypothesis that firms with a high level of concentration may achieve
higher levels of efficiency. However, the hypothesis that revenue diversification has a negative
impact on growth is not satisfied: they do not find a significant relationship between revenue.
II. Research Methodology
Although Frumkin & Keating (2011) suggest that revenue concentration has no
systematic effect on growth, this master thesis investigates the relationship between
diversification and growth again. Frumkin and Keating (2011) state that by concentrating, non-
profit managers have a position to develop specialized skills, but this can be countered by the
argument that this has nothing to do with revenue concentration, but with the human resources
and their capacity and knowledge concerning each revenue stream. A team, large enough, may
develop better skills on several funding streams then a single person may on his a single stream.
With the result that, these revenue streams together may lead to growth of the total revenue
stream, the organization and the program expenses. Therefore, the relationship between
diversification and growth for non-profit organizations will be tested. Besides, Frumkin &
Keating (2011) base their results on the differences between means and medians respectively, of
the most concentrated and most diversified firms of their database. Here, the relationship will be
tested by way of a regression model including several control variables.
The regression model results from two theoretical notions: the concept of Andrés et al.
(2014) will be combined with the concept developed by Carroll and Stater (2008). The first one
18
derived empirical evidence about the relation between diversification and growth in the
for-profit world, the second one investigates whether revenue diversification leads to financial
stability for non-profit firms. By combining both empirical models, the relation between
diversification and growth in the non-profit world can be tested. This brings the following
equation (1):
𝐺𝑖𝑡 = 𝛼𝑖 + 𝛽𝐷𝐼𝑉𝐷𝐼𝑉𝐸𝑅𝑖𝑡−1 + 𝛽𝐷𝐼𝑉2𝐷𝐼𝑉𝐸𝑅𝑖𝑡−12 + 𝛽𝑂𝐸𝑂𝐸𝑖𝑡−1 + 𝛽𝐹𝐹𝐹𝐹𝑖𝑡−1
+ 𝛽𝑆𝑖𝑧𝑒𝑆𝐼𝑍𝐸𝑖,𝑡−1 + 𝛽𝑠𝑒𝑐𝑆𝐸𝐶𝑇𝑂𝑅𝑖𝑡 + 𝛽𝑦𝑌𝐸𝐴𝑅𝑖𝑡 + 𝑢𝑖𝑡 (1)
where i indicates firm i, t indicates the year of observation, αi and βx are the coefficients to be
estimated and uit is the error-term. To control for the endogenous simultaneity problem, 1-year
lag for each independent variable is included. This model is a panel data model, which is
estimated with a fixed effect. Using this panel data model, it is possible to observe the same units
(organizations) collected over a number of periods (years) and thereby it controls for individual
heterogeneity (Torres-Reyna, n.d.). A linear regression model, indexed for both units and time (i
and t) imposes that the intercept term α and the slope coefficients in β are identical for all units
and time periods (de Jong, 2012). The error term varies over units and time and captures all
unobservable factors that affect the dependent variable. Since the same units are repeatedly
observed, it is not representative to assume that the error terms from different periods are
uncorrelated. Therefor a fixed effect is included in the model, which takes care of the
dependence of the error terms and thereby it can capture unobserved individual effects (de Jong,
2012). The fixed effect model removes the effect of the time-invariant characteristics from the
predictor variable, so the predictors’ net effect can be estimated. Thereby, it is assumed that these
time invariant characteristics are unique to the individual and are not correlated with other
individual characteristics (Torres-Reyna, n.d.). Since it is not assumed that the variation across
entities is random and uncorrelated with the predictor, I chose to use the fixed effect model
instead of the random effect model (Torres-Reyna, n.d.).
The assumptions that are made to verify the described model, are mostly innocuous
assumptions, i.e. they are plausible. However, in the fixed effect model, it is assumed that the β’s
are identical across groups, and so the regression estimator reports the average effect. If this
assumption does not hold, this is more problematic. It has to be noticed that this assumption may
19
not hold, since there may be differences across sectors2 (the several sectors are described in
Section III).
The dependent variable, Growth (G) is predicted by three different measurements, the
annual growth in total revenue (TR), program expenses (PE) and fixed assets (FA), (Frumkin &
Keating, 2011). The degree of revenue diversification (DIVER) is measured by a diversification
index which is based on the Herfindahl-Hirschman Index (HERF) (Hirschman, 1964). The
Diversification Index is calculated with the following equation (2):
𝐷𝐼𝑉𝐸𝑅 = 1 − ∑ 𝑝𝑠2𝑛
𝑠=1 (2)
where n is the number of a firm’s revenue sources, and Ps the proportion of the firm’s revenue
from source s. This index is positively related to diversification and will, in principle, always be
between 0 (concentration) and 1 (diversification). Since Andrés et al. (2014) find a quadratic
relation between diversification and the growth opportunities, the quadratic term of DIVER is
also included in the model (DIVER2). In line with the conclusion of Andrés et al. (2014), it is
expected that diversification has an effect on the growth of the organization (βDIV and/or βDIV2 ≠
0). In spite of the expected costs that come with diversification, it is expected that the net result
of the costs and the benefits is positive and therefor leads to growth.
In line with Carroll and Stater (2008), the variables Organizational Efficiency (OE) and
Financial Flexibility (FF) are included in the model. Organizational Efficiency will be measured
by the ratio of administrative and fundraising expenses to total expenses. It is believed that well
organized organizations will receive more revenue in the end and therefor Organizational
Efficiency has a positive relationship with growth (βOE > 0). Financial Flexibility will be
measured with two different proxies which are typically used to measure the financial condition
of a non-profit organization (Jegers and Verschueren, 2006): (1) Debt margin, which provides
information about the extent to which an organization is capable to meet its financial obligations
and is calculated as by dividing the organization’s year-end liabilities by the year-end assets, and
2 The results show that there are, indeed, differences across sectors. When different regressions are ran (not included
in this paper) for each of the different sectors, there are quit some differences observable between the different
βDIV’s; the β’s differ from -8.34 to 7.89. Thereby it has to be noticed that not all the sectors contain enough firm to
build a large enough dataset per sector to get significant results. Therefore, the decision is made to stay with the
fixed effect model as described.
20
(2) Total margin, which gives information about the profitability or increasing value of an
organization, calculated as the proportion of net assets to total revenue. Greater values of Debt
margin indicate less financial flexibility, while greater values of Total margin indicate greater
financial flexibility (Carroll and Stater, 2008). Since financial flexibility contributes to a
healthier organization, it is expected that this has a positive effect on growth (βFF > 0).
Finally, in line with common literature, several control variables are included. Similar to
Andrés et al. (2014) and Carroll and Stater (2008), I control for firm size (SIZE), estimated by
the natural logarithm of the book value of total assets, and for several charity sectors (like health
care, international aid etc.), by including dummies for each mission (SECTOR). Finally, there is
controlled for time-effects by including year dummies (YEAR).
III. Data
In order to examine the relationship between diversification and growth and to investigate
the influence from the other variables, a dataset with financial data of non-profit firms is
collected. This dataset is based on financial data of the Central Bureau on Fundraising (CBF).
The CBF is a Dutch independent foundation who collects and provides information about Dutch
fundraising organizations in the Netherlands. The information includes data from the financial
statements of the organizations. From this financial data, all the variables are derived.
The data consist of annual financial information for each individual non-profit during the
8-year time period between 2005 and 2012. The dataset includes 1,282 different organizations.
Although the majority of organizations (548 or 45.27%) are observed in every year the analysis,
the number of organizations observed each year varies. 132 organizations (10.23%) are observed
for only one year during the time period. The CBF categorizes the Dutch Charity Organizations
in different sectors, related their (main) mission. There are 4 main sectors and 15 sub sectors.
The distribution of the organizations is shown in table 1.
21
Note. The ‘Percentage of total revenue’ refers to the average
percentage that the organizations in the dataset receive from each
revenue source over the time period 2005-2012. On average, the
total income per year was equal to € 3.506.473.648
Table 1
Distribution of organization in several sectors
Sector Mission description # firms % of total
International aid Development aid 533 41.58%
Victim 17 1.33%
Refugee aid 7 0.55%
Health care Health care 130 10.14%
Disabled 70 5.46%
Blind and visually impaired, hard of hearing 13 1.01%
Welfare Community and social goals 219 17.08%
Human rights 20 1.56%
Art and Culture 46 3.59%
Sports and recreation 15 1.17%
Education and Science 41 3.20%
Religion 47 3.67%
Nature and
environment
Environmental interests 27 2.11%
Nature Protection 43 3.35%
Animals 54 4.21%
1282 100%
The CBF collects the financial data of the charity organization since the year 2005. In the first
years, only the information about income and expenses were collected. As from 2008, also
the balance sheet data is gathered. Therefore, some data required to calculate certain variables, is
only available as from 2008. The diversification index, calculated by equation (2), is derived
from the revenue sources each organization uses. The CBF marks 10 different revenue sources,
which are displayed in table 2.
Table 2
Revenue Sources
Revenue source Percentage of
total revenue
Collections 1,75%
Mailings 2,02%
Legacies 6,71%
Gifts, grants, donations and contributions 25,16%
Sales, own lotteries contests 1,98%
Revenue from Joint actions 2,10%
Revenue from actions from third parties 12,13%
Government subsidies 37,51%
Interest income and income from investments 2,71%
Other revenues 7,92%
22
Note. N refers to the total firm-year observations and is equal to 10,217 (i = 1,282; t = 8). Negative
values of DIVER are allowed.
IV. Descriptive statistics
Table 3 provides the descriptive statistics of all variables, including the control variables,
included in the model. The number of observations (N) differs strongly as a result of the absence
of the balance sheet data from 2005 to 2007. The diversification Index (DIVER) should, in
principle, always lie between 0 and 1. This differs in the dataset since it occurs that Dutch charity
organizations have a negative income from the revenue source ‘Interest income and income from
investments’. This negative income may lead to a negative diversification index, which is the
case in 95 observations. This negative indexes cause a lower mean value of DIVER (0.2370).
When the negative outcomes are excluded from the database, the mean value of DIVER
increases to 0.3391 (Appendix A). Logically, removing the 95 negative observations has also
impact on the standard deviation of the Diversification Index: the standard deviation decreases
from 3.8682 (table 3) to 0.2408 (Appendix A). The negative values also explain the maximum of
the extreme high quadratic term of the diversification index (DIVER2). When the negative values
are removed, the mean and standard deviation of DIVER2 are 0.1730 and 0.1702 respectively
(Appendix A). The maximum will lie on 0.6571 (Appendix A).
Table 3
Summery statistics for the full sample (2005-2012)
Variable N Mean
Standard
Deviation Min. Max.
Growth Total Revenue (TR) 5833 0.4026 4.2145 -0.9865 185.0411
Growth Program Expenses (PE) 5765 0.5340 6.5703 -0.9988 302.4643
Growth Fixed Assets (FA) 1813 2.4082 32.1507 -1.0000 815.5970
DIVER 7269 0.2370 3.8682 -210.8668 0.8106
DIVER2 7269 15.0174 724.0323 0 44,464.8000
Organizational Efficiency (OE) 4921 0.1057 0.1555 0.0001 1
Debt Margin (DM) 2773 0.4179 1.7426 0.0000 70.2558
Total Margin (TM) 4707 2.2582 10.6417 0.0021 583.4288
Size 4747 12.4945 2.5016 4.5468 19.5310
When running the regression, the different sectors that the CBF identifies (table 1), will
be used as a control variable. In Appendix B, the mean value of each variable is reported per
sector. Due to heterogeneity, there are quite some differences between the sectors, both in the
23
growth variables, the diversification index and in the other variables. From the mean values
reported in Appendix B, it does not follow directly that the sector with the higher diversification
index is also the sector who experiences the highest growth.
Table 3 shows the diversity of the growth variables - total revenues, program expenses
and fixed assets. All three know small negative values and big positive values. The mean value
of Growth in Fixed Assets lies higher than the mean value of Growth in Total Revenue and
Growth in Program Expenses, but also the standard deviation is higher for this latter. Tables 5 to
7 report the statistics of each dependent and independent variable per level of growth. For each
one of them, the 5% or higher growth level knows high variation: the standard deviations are
high and also the mean values lie at an extreme higher value, while the N for each of this level is
much lower compared to the other levels.
In table 4, an overview of the mean growth per level of diversification is given. It is
expected that for higher levels of diversification, the organization will experience higher growth.
The numbers reported in table 4 suggest that this expectation cannot be confirmed. For each of
the three growth measurements, a parabolic trend is present: when one increases his level of
diversification from zero or less to the first level (0 to 0.25), the growth of the firm increases.
When a firm reaches higher levels of diversification, this effect turns around and one experience
less growth. This suggestion confirms the addition of the quadratic term of DIVER in the model.
Table 4
Diversification Index
Diversification
Index Freq. % of total
Cum. Rel.
freq Growth PE Growth TR Growth FA
Less than 0 528 7.26% 7.26% 0.6442 0.3083 -0.0342
0 to 0.25 2,436 33.51% 40.78% 0.7471 0.6140 2.0198
0.25 to 0.5 2,067 28.44% 69.21% 0.6551 0.4138 1.3201
0.5 to 0.75 2,095 28.82% 98.03% 0.2074 0.1930 3.5170
0.75 to 1 143 1.97% 100% 0.0248 0.1283 2.6084
Total 7,269 100% 0.5354 0.4026 2.4191
This effect can also be shown the other way around. For each of the dependent variables,
growth in Program Expenses, Total Revenue and Fixed Assets, a table is created to oversee the
mean values of each independent variable for each level of growth (table 5 to 7).
Table 5
Growth in Total Revenue
Growth in Total
Revenue Freq. % of total
Cum. Rel.
freq. Mean SD Diver Diver2 OE
Debt
Margin
Total
Margin Size
Less then - 0.5% 379 6.5 6.5 -0.6809 0.1270 -1.5507 284.6278 0.1654 0.2894 9.1800 11.2700
-0.5% to 0 2,248 38.54 45.04 -0.1825 0.1344 0.3349 0.1766 0.0943 0.4583 1.9673 12.6060
0 to +0.5% 2,277 39.04 84.07 0.1726 0.1322 0.3721 0.1991 0.0839 0.4376 1.3989 13.2125
+0.5% to +1% 433 7.42 91.5 0.7013 0.1455 0.3428 0.1783 0.0985 0.2678 1.5683 12.1629
+1% to +5% 424 7.27 98.77 2.0439 1.0332 0.2881 0.3173 0.1043 0.3208 1.7944 11.6403
+5% or higher 72 1.23 100 20.1874 31.9198 0.2198 0.0950 0.0750 0.3561 1.3755 11.4083
Total 5,833 100 0.4026 4.2145 0.2227 18.6770 0.0958 0.4226 2.2033 12.6429
Table 6
Growth in Program Expenses
Growth in
Program Expenses Freq. % of total
Cum. Rel.
freq. Mean SD Diver Diver2 OE
Debt
Margin
Total
Margin Size
Less then - 0.5% 363 6.3 6.3 -0.6817 0.1321 -0.3580 123.1511 0.1887 0.3193 3.3951 11.1935
-0.5% to 0 2,148 37.26 43.56 -0.1759 0.1349 0.1530 29.4276 0.0894 0.3799 2.0706 12.7454
0 to +0.5% 2,350 40.76 84.32 0.1679 0.1269 0.3522 0.5566 0.0760 0.4892 1.6058 13.2804
+0.5% to +1% 421 7.3 91.62 0.7044 0.1506 0.3134 0.1777 0.0701 0.3375 1.9630 12.1486
+1% to +5% 392 6.8 98.42 2.0498 0.9537 0.2764 0.1302 0.0723 0.2550 2.3512 11.5086
+5% or higher 91 1.58 100 24.2797 46.4539 0.2321 0.1020 0.0806 0.2762 1.2462 10.5336
Total 5,765 100 0.5340 6.5703 0.2233 18.9748 0.0875 0.4195 1.9606 12.7136
Table 7
Growth in Fixed Assets
Growth in Fixed
Assets Freq. % of total
Cum. Rel.
freq. Mean SD Diver Diver2 OE
Debt
Margin
Total
Margin Size
Less then - 0.5% 125 6.89 6.89 -0.6678 0.1473 0.4247 0.2289 0.1031 0.4928 1.9884 13.0726
-0.5% to 0 1,011 55.76 62.66 -0.1722 0.1385 0.4029 0.3698 0.1035 0.3713 2.7213 13.9047
0 to +0.5% 426 23.5 86.16 0.1445 0.1286 0.4597 0.2729 0.0890 0.4332 2.2653 15.0585
+0.5% to +1% 75 4.14 90.29 0.7111 0.1498 0.4287 0.2352 0.1048 0.2684 3.0138 13.9582
+1% to +5% 116 6.4 96.69 2.2165 1.0038 0.2758 2.5181 0.1022 0.3257 1.3006 14.0450
+5% or higher 60 3.31 100 70.8615 163.7166 0.4827 0.2787 0.0972 0.2122 1.8127 14.3861
Total 1,813 100 2.4082 32.1507 0.4134 0.4657 0.0997 0.3820 2.4543 14.1456
In each of these tables, especially in table 5 (Total Revenue) and table 6 (Program Expenses) the
same effect can be observed. For small levels of growth, the Diversification Index (DIVER) is
small or even negative, for higher values of growth the Diversification Index increases, but for
the highest values the index decreases.
Appendix C shows the correlation matrix of each of the variables. This matrix in does not
show a significant relation between the diversification Index and the Growth variables.
Appendix D shows the correlation matrix from the variables as included in equation (1), with a
1-year lag for the dependent variables. This matrix shows some significant correlations, for
example between the Diversification Index and the Growth in Total Revenue. But, correlation
and causality are two very different things and therefore there cannot anything be concluded
from these correlation matrices. The next sector gives the results of running the regressions,
which, in contrast, can tell about causality.
V. Results
To determine whether fixed effect panel data model or a random effect panel data model
is preferred, a Hausman test should be executed. This test tests whether the unique errors are
correlated with the regressors. The null hypothesis (H0) of this test states they are not and that the
preferred model is the random effect model (Torres-Reyna, n.d.). For both Growth in Program
Expenses and Growth in Total Revenue as dependent variable, the P-value of the Hausman test
turns out to be 0.000 and so the most efficient model turns out to be the fixed effects model. For
Growth in Fixed Assets the P-value is equal to 0.7737, which suggest that the Random Effects
model should be more efficient. The fixed effect model will be applied to all of the three
different dependent variables. Since the null hypothesis cannot be rejected for Growth in Fixed
Assets, the Random Effects model will also be investigated for this dependent variable.
Tables 8, 9 and 10 present the fixed effects regression results from equation (1) for each
of the different dependent variables; the growth in Total Revenue, Program Expenses and Fixed
Assets respectively. In these tables, various specifications of the equation are estimated. Each of
the specifications are explained in the notes beneath the tables. The results from the panel data
model with random effects are shown in table 11.
26
Note. This table provides the fixed effects regression results of equation (1) with growth in Total Revenue as
dependent variable. Specification (a) regresses only the diversification index (DIVER) on the dependent variable,
controlled for size and year. Specification (b) includes also the quadratic term of DIVER. In specification (c) and
(d) also the variables Organizational Efficiency (OE) and a proxy for Financial Flexibility (FF) are included (Debt
Margin or Total Margin respectively). In specification (a)-(d) the negative values of DIVER are included.
Specification (e) contains the same variables as specification (c), but before running this regression, the negative
values of DIVER are removed. Obs refers to the firm-year observations for each regression, the total observations
are equal to 10,217 (i = 1,283; t = 8); the F statistic is used to test whether all coefficients are different from zero
(H0: β1 = … = βi = 0); R square refers to the part of variation in the dependent variable explained by the model. The
standard error is shown in parenthesis under the coefficients. The ***, ** and * denote statistical significance at
the 1%, 5% and 10% level, respectively.
Table 8
Regression results Growth in Total Revenue
Dependent variable:
Growth Total Revenue
Variable (a) (b) (c) (d) (e)
DIVER -0.03665***
(0.0130)
-0.3454***
(0.0718)
-0.3589***
(0.1228)
-0.7662***
(0.1913)
-0.3745***
(0.1331)
DIVER2 -0.0018***
(0.0004)
-0.0020***
(0.0007)
-0.0184**
(0.0080)
-0.0021***
(0.0008)
Organizational Efficiency (OE) 2.4124***
(0.7904)
4.8597***
(0.6332)
2.3759***
(0.7866)
Total Margin (TM) 0.0378***
(0.0078)
0.0375***
(0.0077)
Debt Margin (DM) -0.0423
(0.0380)
Size -2.3293***
(0.1426)
-2.3266***
(0.1421)
-2.5344***
(0.1668)
-1.2642***
(0.1298)
-2.4339***
(0.1682)
Year 2010 0.1360
(0.1463)
0.1763
(0.1460)
0.1801
(0.1720)
-0.1370
(0.0962)
0.2192
(0.1715)
2011 -0.0544
(0.1480)
-0.0139
(0.1477)
-0.0052
(0.1740)
-0.1857*
(0.0978)
0.0266
(0.1733)
2012 0.0511
(0.1509)
0.0901
(0.1505)
0.0953
(0.1770)
-0.1075
(0.1095)
0.1215
(0.1765)
Constant 29.7159***
(1.7989)
29.7508***
(1.7838)
32.2206***
(2.1117)
17.5818***
(1.7883)
30.9367***
(2.1316)
Obs 3413 3413 2912 1967 2892
F stat 57.1317*** 51.1756*** 36.5353*** 27.0517*** 33.5534***
R square 0.1127 0.1202 0.1350 0.1464 0.1262
27
Note. This table provides the fixed effects regression results of equation (1) with growth in Program Expenses
as dependent variable. Specification (a) regresses only the diversification index (DIVER) on the dependent
variable, controlled for size and year. Specification (b) includes also the quadratic term of DIVER. In
specification (c) and (d) also the variables Organizational Efficiency (OE) and a proxy for Financial Flexibility
(FF) are included (Debt Margin or Total Margin respectively). In specification (a)-(d) the negative values of
DIVER are included. Specification (e) contains the same variables as specification (c), but before running this
regression, the negative values of DIVER are removed. Obs refers to the firm-year observations for each
regression, the total observations are equal to 10,217 (i = 1,283; t = 8); the F statistic is used to test whether all
coefficients are different from zero (H0: β1 = ... = βi = 0); R square refers to the part of variation in the dependent
variable explained by the model. The standard error is shown in parenthesis under the coefficients. The ***, **
and * denote statistical significance at the 1%, 5% and 10% level, respectively.
Table 9
Regression results Growth Program Expenses
Dependent variable:
Growth Program Expenses
Variable (a) (b) (c) (d) (e)
DIVER -0.0082
(0.0309)
-0.0832
(0.1718)
0.0020
(0.2629)
0.0397
(0.3350)
-0.0809
(0.2919)
DIVER2 -0.0005
(0.0010)
-0.0003
(0.0015)
0.0022
(0.0139)
0.0010
(0.0018)
Organizational Efficiency (OE) 53.4422***
(2.3040)
25.6913***
(1.2800)
53.6592***
(2.3189)
Total Margin (TM) 0.1178*
(0.0658)
0.1067
(0.0679)
Debt Margin (DM) -0.0062
(0.0630)
Size -1.6833***
(0.3512)
-1.6825***
(0.3454)
-1.8674***
(0.3572)
0.1272
(0.2158)
-2.4339***
(0.1682)
Year 2010 0.1795
(0.3511)
0.1913
(0.3519)
0.1042
(0.3681)
0.1644
(0.1606)
0.2192
(0.1715)
2011 -0.2638
(0.3547)
-0.2520
(0.3555)
-0.3717
(0.3712)
-0.0855
(0.1819)
0.0267
(0.1733)
2012 -0.4266
(0.3617)
-0.4152
(0.3623)
-0.3139
(0.3784)
-0.0855
(0.1819)
0.1215
(0.1765)
Constant 21.9628***
(4.3533)
21.87389***
(4.3540)
19.6380***
(4.5340)
-3.8102
(2.9763)
30.9367***
(2.1316)
N 3353 3353 2859 1852 2839
F stat 5.8988*** 4.9628*** 73.0794*** 50.8665*** 72.5131***
R square 0.0131 0.0133 0.2410 0.2457 0.2411
28
Note. This table provides the fixed effects regression results of equation (1) with growth in Fixed Assets as
dependent variable. Specification (a) regresses only the diversification index (DIVER) on the dependent
variable, controlled for size and year. Specification (b) includes also the quadratic term of DIVER. In
specification (c) and (d) also the variables Organizational Efficiency (OE) and a proxy for Financial Flexibility
(FF) are included (Debt Margin or Total Margin respectively). In specification (a)-(d) the negative values of
DIVER are included. Specification (e) contains the same variables as specification (c), but before running this
regression, the negative values of DIVER are removed. Obs refers to the firm-year observations for each
regression, the total observations are equal to 10,217 (i = 1,283; t = 8); the F statistic is used to test whether all
coefficients are different from zero (H0: β1 = … = βi = 0); R square refers to the part of variation in the
dependent variable explained by the model. The standard error is shown in parenthesis under the coefficients.
The ***, ** and * denote statistical significance at the 1%, 5% and 10% level, respectively.
Table 10
Regression results Growth Fixed Assets (Fixed Effects)
Dependent variable:
Growth Fixed Assets
Variable (a) (b) (c) (d) (e)
DIVER -0.5315
(0.7045)
-3.5027
(2.3465)
-5.5548
(3.6659)
-6.0583
(3.9989)
-5.7587
(3.7435)
DIVER2 -0.0825
(0.0621)
-0.1515
(0.1529)
-0.1720
(0.1661)
-0.1596
(0.1559)
Organizational Efficiency (OE) -2.3522
(15.8039)
-1.9224
(19.6608)
-2.5769
(15.9673)
Total Margin (TM) -0.0156
(0.0756)
-0.0155
(0.0760)
Debt Margin (DM) -1.2552
(2.4901)
Size -3.2037
(2.9214)
-3.0919
(2.9217)
-3.2792
(2.6206)
-5.3517
(3.6057)
-3.6540
(2.7618)
Year 2010 -4.1115*
(2.3456)
-3.9257*
(2.3490)
-2.2211
(2.1718)
-2.3457
(2.3063)
-2.2395
(2.1924)
2011 -4.0485*
(2.3945)
-3.8603
(2.3980)
-2.0278
(2.2032)
-1.8155
(2.3541)
-2.0403
(2.2237)
2012 -2.1158
(2.4590)
-2.0220
(2.4593)
0.1611
(2.2588)
0.3433
(2.4118)
0.2076
(2.2817)
Constant 50.4255
(40.9939)
49.9747
(40.9823)
52.4409
(37.2581)
83.9867
(52.5188)
57.9678
(39.3285)
N 1796 1796 1566 1390 1554
F stat 1.2989 1.3743 1.1439 .2258 1.1744
R square 0.0054 0.0068 0.0088 0.0102 0.0091
29
Note. This table provides the random effects regression results of equation (1) with growth in Program Fixed
Assets as dependent variable. Specification (a) regresses only the diversification index (DIVER) on the
dependent variable, controlled for size and year. Specification (b) includes also the quadratic term of DIVER. In
specification (c) and (d) also the variables Organizational Efficiency (OE) and a proxy for Financial Flexibility
(FF) are included (Debt Margin or Total Margin respectively). In specification (a)-(d) the negative values of
DIVER are included. Specification (e) contains the same variables as specification (c), but before running this
regression, the negative values of DIVER are removed. Obs refers to the firm-year observations for each
regression, the total observations are equal to 10,217 (i = 1,283; t = 8); the Chi2 statistic is used to test whether
all coefficients are different from zero (H0: β1 = ... = βi = 0); R square refers to the part of variation in the
dependent variable explained by the model. The standard error is shown in parenthesis under the coefficients.
The ***, ** and * denote statistical significance at the 1%, 5% and 10% level, respectively.
Table 11
Regression results Growth Fixed Assets (Random Effects)
Dependent variable:
Growth Fixed Assets
Variable (a) (b) (c) (d) (e)
DIVER -0.3819
(0.5952)
-1.8639
(1.7677)
-2.7194
(24680)
-1.4177
(2.4680)
-2.7555
(2.4646)
DIVER2 -0.0422
(0.0473)
-0.0402
(0.1046)
0.0108
(0.1052)
-0.0417
(0.1052)
Organizational Efficiency (OE) 2.7484
(7.1079)
1.1674
(8.1410)
2.7904
(7.1294)
Total Margin (TM) 0.0055
(0.0489)
0.0056
(0.0491)
Debt Margin (DM) -0.4981
(1.2900)
Size 0.0372
(0.3582)
0.0842
(0.3621)
0.0384
(0.3754)
0.0522
(0.3903)
0.0267
(0.3774)
Year 2010 -3.9304*
(2.2210)
-3.8602*
(2.2226)
-2.3004
(2.0338)
-2.5642
(2.1687)
-2.3218
(2.0534)
2011 -3.4410
(2.1868)
-3.3769
(2.1881)
-1.9509
(2.0025)
-2.8052
(2.1266)
-1.9778
(2.0211)
2012 -2.6598
(2.2002)
-2.6521
(2.2003)
-0.9049
(2.0089)
-0.5572
(2.1967)
-0.9221
(2.0277)
Constant 4.6302
(5.3045)
4.5544
(5.3055)
4.3742
(5.6568)
3.5401
(6.0379)
4.6000
(5.7049)
Obs 1796 1796 1566 1390 1554
Chi2 4.30 5.09 6.14 6.27 6.14
R square 0.0042 0.0054 0.0064 0.0062 0.0065
30
The Diversification Index turns out to have a statistically significant negative relation
with Growth in Total Revenue (table 8). This negative relationship is observable in each of the
specifications. Compared to the simple regression in specification (a), where only SIZE and
YEAR are included in the model, this effect is stronger when the quadratic term of DIVER is
included (DIVER2, specification (b)) and becomes even more stronger when the other
independent variables Organizational Efficiency (OE) and Financial Flexibility (FF) are included
in the model (specification (c)). Since an increase of 1 in DIVER is not realistic, it is necessary to
interpret the estimator of βdiv correctly. If one increases his diversification level with 0.10 (ceteris
paribus), he will experience on average, a negative growth – a shrink – of almost 3.6%
(specification (c)), according to this model. Thereby, it has to be noticed that the value of DIVER,
is in principle always between 0 and 1. The relationship between the Diversification Index and
Growth in Total Revenue turns out to be a quadratic, mountain shaped relation. This quadratic
relation is statistically significant, but minimal: for specification (b) and (c) the effect is very
small and barely observable since the accompanying estimator is close to zero (-0.0018 and
-0.0020 for specification (b) and (c) respectively). Since DIVER – and therefor DIVER2 – can, in
principle, only lie between 0 and 1, the influence of this quadratic term is negligible. The
economic significance of this quadratic term can therefore be rejected. The other independent
variables (OE and FF) turn out to have also an effect on growth. Organizational Efficiency turns
out to have a positive significant effect, with a statistically significant coefficient estimator.
Notwithstanding that this estimator has to be placed in perspective since the values of
Organizational Efficiency will always lie between 0 and 1, the economic significance is quite
large: if one achieves a 10% higher level of Organizational Efficiency (ceteris paribus), he will
experience, on average, a 24% growth of Total Revenue. With respect to Financial Flexibility,
the different proxies show different results. When Total Margin (TM) is used as a proxy
(specification (c)), there is a statistically significant positive effect. The economic significance of
this estimator is quite large, since an increase of one standard deviation (10.64) of the Total
Margin Ratio will lead, on average and ceteris paribus, to an increase of the growth in Total
Revenue of 40%. However, when Debt Margin (DM) is included in the model instead of Total
Margin (specification (d)), there is no statistical significant effect. This might be due to the
amount of observations, which is in specification (d) much lower than in the other specifications
as a result of the lack of balance sheet data, which is necessary to calculate the debt margin. This
31
lower amount of observations might also be the reason for the differences between the other
estimators of specification (d) and the further specifications. Specification (e) is estimated from a
database where the negative values of DIVER are excluded. In this specification are the same
independent variables used as in specification (c), since this latter show the most useful and
complete outcomes (comparing specifications (a) to (d)). The inclusion or the exclusion of the
negative values do not have that much effect; the differences between the specifications (c) and
(e) are minimal. The control variable SIZE does have a statistical significant effect in all
specifications, unlike YEAR, which does not have any significant effect on growth in Total
Revenue. For all the specifications, the included variables turn out to be statistically jointly
significant since the F-statistics are significant and therefor the null hypothesis can be rejected
(H0: all β’s are equal to zero). The specification explain all about 13% of the overall variation of
in the Growth of Total Revenue (R-square, specification (c)).
When the regressions are ran with Growth in Program Expenses as dependent variable
(table 9), the results are not that clear as for Total Revenue (table 8). The estimation coefficients
of DIVER and DIVER2 do not show any statistically significant values, not in any of the
specifications. There cannot be drawn any conclusion about the relation between the
Diversification Index and the Growth in Program Expenses, based on this model and database.
Organizational Efficiency does have, again, a strong positive relationship with the dependent
variable: this effect is even much stronger for the growth in Program Expenses with respect to
growth in Total Revenue, given the size of the estimator. The results suggest that when
estimating the relationship of Financial Flexibility and Growth in Program Expenses with Total
Margin, there is a significant positive effect (specification (c)). This is actually not confirmed
when the negative values of DIVER are removed from the database (specification (e)). The
control variable SIZE has again a significant positive effect, and YEAR does not. Again, the
included variables of each of the specifications turn out to be jointly significant and useful, since
the F statistics, for each specification, is large enough to reject the null hypothesis (H0: all β’s are
equal to zero). For specification (c), the model explains about 24% of the variation in the growth
of the Program Expenses (R-square).
The fixed effects model with Growth in Fixed Assets as dependent variable, turns out to
have not any significant results (table 10). Not for the individual estimators, but also the included
32
variables are not jointly significant (due to the small F statistic). This might be due to the lack of
observations; since the Growth in Fixed Assets is calculated by balance sheet items, there is
limited data about this variable available. Unfortunately, there cannot be drawn any conclusion
about the relation of the diversification index with the Growth in Fixed Assets. Also when the
same model is estimated with random effects (table 11), there are no significant results.
VI. Conclusions
The results presented in the previous section suggest, like expected, a relation between
revenue diversification and growth. This supports the hypothesis stated in section II (βDIV and/or
βDIV2 ≠ 0). Nevertheless, it was expected that there was a positive relationship, but these results
suggest a negative one: the data shows that when an organization is more diversified, he will, on
average, experience a negative growth. The argument – stated in section IV – that a team large
enough, with human resources and knowledge about every different revenue stream, can develop
skills to be even more effective at fundraising does not hold. Even when the costs of
diversification are ignored and only the growth in Total Revenue is measured, diversification has
a negative impact. Focusing on several particular revenue streams, seems to be the right strategy
to achieve growth.
These results confirm the expectation of Frumkin and Keating (2011), which suggested a
positive relation between revenue concentration and growth, i.e. a negative relation between
diversification and growth. Although they did not find this relationship confirmed, the results
presented here suggest that they were right. Their statement seems to be valid: non-profits can
create long-term marketing relationships and a reputation, by developing specialized skills that
enable effective fundraising. The results are also in line with the first wave of research about the
question whether (corporate) diversification leads to the creation or destruction of firm value
(Martin and Sayrak, 2003): the results suggest that diversification leads to value destruction since
the Growth in Total Revenue decreases when one diversifies. The statement that the
diversification discount disappears (or drops) when one controls for several firm factors (Campa
and Kedia, 2002) is not confirmed by the results: when controlled for other factors like Financial
Flexibility and Organizational Efficiency, the results become more significant and show a larger
negative effect. Kingma (1993) pointed out that non-profit organizations bear the additional risk
33
of not-receiving any addition expected funding at all. This might indicate the different
conclusions between the literature which confirms the diversification premium and the
conclusion presented here: diversification may lead to a diversification premium for for-profit
firms, however not for non-profit organizations. The factors Organizational Efficiency and
Financial Flexibility have, in line with the hypotheses (βOE > 0 and βFF > 0), a significant effect
on Growth. Organizational Efficiency turns out to have a positive effect on Growth, which is in
line with the expectations; being organized well has a good impact on Growth. Financial
Flexibility, with Total Margin as proxy, has also a positive impact on Growth, which confirms
the expectation.
The results are different from the results of Andrés et al. (2014), who do suggest a
U-shaped quadratic relation between diversification and growth. The estimator of the quadratic
term is converted (U-shaped vs. mountain shaped), but meanwhile, the estimator of the
Diversification Index shows the same negative significant effect. Andrés et al. (2014) interpret
the U-shaped effect by suggesting that in early stages of diversification, diversification hinders
growth, but in later stages, it becomes a source of further growth options. It seems that this last
stage is different for non-profit firms (Andrés et al. (2014) investigate the relationship for
for-profit organizations), which is in line with the suggestion about the diversification discount
above. The differences between the work of Andrés et al. (2014) and this master thesis may also
be (partly) caused by the different proxies used; Andrés et al. (2014) use the market to book asset
ratio, Tobin’s Q and the ratio of R&D expenses to total sales as a proxies for Growth. All these
proxies are not applicable to non-profit firms since there is no market data available, and the
R&D expenses for non-profit firms are of a different nature. The difference in growth
measurement that are used may cause the different results.
Concluding, the main question stated in the title and introduction of this master thesis –
Revenue Diversification in Dutch Charity Organizations: Does it lead to Growth? – should get a
negative answer. The results indicate a negative relation which means that organizations who
increase their level of diversification, experience – on average and ceteris paribus – a shrinkage
of their total revenues. Revenue diversification does not lead to growth.
34
VII. Discussion and recommendations
The study presented here can be improved. This section provides some suggestions which
one can keep in mind by investigating the same or a related topic.
First, the results might be improved by redoing this study with a broader database. Due to
a lack in balance sheet data, the number of observations for some variables were small, which
might lead to less or in-significant results. When a database is constructed with more years of
observations, those results might get more significant. Second, these database only includes
Charity organizations from the Netherlands. Although there are enough organizations to
investigate, a broader study might be interesting to also investigate the situation in other
countries.
Third, Santalo & Becerra (2008) state that the conflicting results and interpretation
between the scholars can be explained since they estimate the effect of diversification over a
large variety of industries. Since this study investigates only one industry (Dutch Charity
organizations), this problem is not applicable. It can be discussed whether it justified to assume
that the average effect across the several sectors is equal. It seems that this assumption does not
hold, like stated in the footnote of section II. With the dataset used, the regressions per sector did
deliver only significant results for a few sectors with a larger amount of organizations. A lot of
sectors contain only a few organizations in this dataset (table 1), and therefore no significant
results can be obtained. A larger dataset can overcome this problem. In addition, Santalo and
Beccera (2008) state that diversified organizations experience a diversification discount only,
when they compete in industries where focused firms hold a considerable market share. It is hard
to argue whether the Dutch non-profit market is a concentrated market or not. Further
investigation of the Dutch charity market in this perspective, might be interesting: to investigate
whether the non-profit market differs from for-profit markets in concepts of soft information and
vertically integration – two concepts named by Santalo and Becerra (2008) to explain the
heterogeneity across sectors – and whether there are differences between the several sectors
within the non-profit sector.
Fourth, it might be that the relationship between Organizational Efficiency and Growth is
also quadratic, since an organization spending too much on the organization itself, might lose the
35
trust of donators, which might lead to less donations (Carroll & Stater, 2009). This effect is not
investigated and might be interesting for further research. Next to that, in response to similar
studies, this master thesis controls for size, measured by the natural logarithm of the book value
of assets. It might be interesting to control also for the size of total revenue, since this might
influence the possibility to grow.
Finally, this master thesis does not investigate the relationship between the growth of the
net result of the charity organization and diversification. Diversification brings several costs, like
personnel costs, administrative costs and fundraising expenses (Kingma, 1993). Since
diversification does not lead to a growth in total revenue, it is expected that when these costs are
also taken into account, the result becomes even more significant. Nevertheless, this might be an
interesting topic for future research.
36
REFERENCES
Berger, P. G., & Ofek, E. (1995). Diversification's effect on firm value. Journal of financial economics, 37(1), 39-
65.
Berger, P. G., & Ofek, E. (1996). Bustup Takeovers of Value‐Destroying Diversified Firms. The Journal of Finance,
51(4), 1175-1200.
Bodie, Z., Kane, A., & Marcus, A. J. (2011). Investments and portfolio management. New York, NY. McGraw-
Hill/Irwin.
Bowman, W. (2002). The uniqueness of nonprofit finance and the decision to borrow. Nonprofit Management and
Leadership, 12(3), 293-311.
Campa, J. M., & Kedia, S. (2002). Explaining the diversification discount. The Journal of Finance, 57(4), 1731-
1762.
Carroll, D. A., & Stater, K. J. (2009). Revenue Diversification in nonprofit organizations: does it lead to financial
stability?. Journal of Public Administration Research and Theory, 19(4), 947-966.
Centraal Bureau voor de Statistiek [CBS] (2014). Nederland in 2013, een economisch overzicht. Retrieved from
http://www.cbs.nl/NR/rdonlyres/911988A7-1223-49D7-BEEC-1C3927924FD1/0/2014Nederlandin2013pub.pdf
Centraal Bureau Fondsenwerving [CBF] (2014). Financiële overzichten. Retrieved from
http://www.cbf.nl/cijfers/baten-en-lasten/
Chevalier, J. (2000). Why do firms undertake diversifying mergers? An analysis of the investment policies of
merging firms. Unpublished working paper. University of Chicago.
Corbin, J. J. (1999). A study of factors influencing the growth of nonprofits in social services. Nonprofit and
Voluntary Sector Quarterly, 28(3), 296-314.
De Andrés, P., de la Fuente, G., & Velasco, P. (2014). Growth opportunities and the effect of corporate
diversification on value. Article in press, The Spanish Review of Financial Economics.
Elton, E. J., & Gruber, M. J. (1997). Modern portfolio theory, 1950 to date. Journal of Banking & Finance, 21(11),
1743-1759.
Erdorf, S., Hartmann-Wendels, T., Heinrichs, N., & Matz, M. (2013). Corporate diversification and firm value: a
survey of recent literature. Financial Markets and Portfolio Management, 27(2), 187-215.
Ferris, S. P., Sen, N., Lim, C. Y., & Yeo, G. H. (2002). Corporate focus versus diversification: the role of growth
opportunities and cashflow. Journal of International Financial Markets, Institutions and Money, 12(3), 231-252.
Friend, I., & Blume, M. (1970). Measurement of portfolio performance under uncertainty. The American Economic
Review, 561-575.
Frumkin, P., & Keating, E. K. (2011). Diversification reconsidered: the risks and rewards of revenue
concentration. Journal of social entrepreneurship, 2(2), 151-164.
Graham, J. R., Lemmon, M. L., & Wolf, J. G. (2002). Does corporate diversification destroy value?. The Journal of
Finance, 57(2), 695-720.
Grossman, D. A. (1992). Paying nonprofits: Streamlining the new york state system. Nonprofit Management and
Leadership, 3(1), 81-91.
Hadlock, C. J., Ryngaert, M., & Thomas, S. (2001). Corporate Structure and Equity Offerings: Are There Benefits to
Diversification?. The journal of business, 74(4), 613-635.
Harris, M., Kriebel, C. H., & Raviv, A. (1982). Asymmetric information, incentives and intrafirm resource
allocation. Management Science, 28(6), 604-620.
Hirschman, A. O. (1964). The paternity of an index. The American Economic Review, 761-762.
Jegers, M. (1997). Portfolio theory and nonprofit financial stability: A comment and extension. Nonprofit and
voluntary sector quarterly, 26(1), 65-72.
37
Jegers, M., & Verschueren, I. (2006). On the capital structure of non‐profit organisations: an empirical study for
californian organisations. Financial Accountability & Management, 22(4), 309-329.
Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and takeovers. The American economic
review, 323-329.
Jensen, M. C., & Murphy, K. J. (1990). Performance pay and top-management incentives. Journal of political
economy, 225-264.
John, K., & Ofek, E. (1995). Asset sales and increase in focus. Journal of Financial Economics, 37(1), 105-126.
Jong, de, F. (2012). Models Based on Panel Data. Retrieved from https://edubb.uvt.nl/bbcswebdav/pid-785197-dt-
content-rid-1880431_1/courses/323063-2013-2014/Paneldata%282%29.pdf
Kingma, B. R. (1993). Portfolio theory and nonprofit financial stability. Nonprofit and Voluntary Sector
Quarterly, 22(2), 105-119.
Lang, L.H.P., Stulz, R.M., (1994). Tobin’s q, corporate diversification and firm performance. Journal of political
Economy 102, 1248-1280
Lee, C. F. (1977). Functional form, skewness effect, and the risk-return relationship. Journal of financial and
quantitative analysis, 12(01), 55-72.
Lewellen, W. G. (1971). A pure financial rationale for the conglomerate merger.The Journal of Finance, 26(2), 521-
537.
Markowitz, H. (1952). Portfolio selection. The journal of finance, 7(1), 77-91.
Martin, J. D., & Sayrak, A. (2003). Corporate diversification and shareholder value: a survey of recent literature.
Journal of Corporate Finance, 9(1), 37-57.
Ramanujam, V., & Varadarajan, P. (1989). Research on corporate diversification: A synthesis. Strategic
management journal, 10(6), 523-551.
Rijksoverheid (2014). Financiering ontwikkelingssamenwerking. Retrieved from
http://www.rijksoverheid.nl/onderwerpen/ontwikkelingssamenwerking/financiering-ontwikkelingssamenwerking
Santalo, J., & Becerra, M. (2008). Competition from specialized firms and the diversification–performance
linkage. The Journal of Finance, 63(2), 851-883.
Sharpe, W. F. (1966). Mutual fund performance. Journal of business, 119-138.
Shleifer, A., & Vishny, R. W. (1989). Management entrenchment: The case of manager-specific
investments. Journal of financial economics, 25(1), 123-139.
Sorensen, E. H., Qian, E., Schoen, R., & Hua, R. (2004). Multiple alpha sources and active management. The
Journal of Portfolio Management, 30(2), 39-45.
Stein, J. C. (1997). Internal capital markets and the competition for corporate resources. The Journal of
Finance, 52(1), 111-133.
Stowe, J. D., & Xing, X. (2006). Can growth opportunities explain the diversification discount?. Journal of
Corporate Finance, 12(4), 783-796.
Tobin, J. (1958). Liquidity preference as behavior towards risk. The review of economic studies, 65-86.
Torres-Reyna, P. (n.d.). Panel Data Analysis, Fixed and Random Effects. Retrieved from
http://dss.princeton.edu/training/Panel101.pdf
Villalonga, B. (2000). Matching BITS to COMPUSTAT: Towards richer data for large sample research within
firms. Unpublished manuscript, Anderson Graduate School of Management, University of California, Los Angeles.
Villalonga, B. (2004). Diversification discount or premium? New evidence from the business information tracking
series. The Journal of Finance, 59(2), 479-506.
38
APPENDICES
Appendix A
Summery statistics for the full sample (2005-2012) when DIVER > 0
Variable N Mean
Standard
Deviation Min. Max.
Growth Fixed Assets (FA) 1798 2.4280 32.2838 -0.1000 815.5971
Growth Total Revenue (TR) 5761 0.4024 4.2213 -0.9865 185.0411
Growth Program Expenses (PE) 5694 0.5356 6.6076 -0.9931 302.4643
DIVER 7174 0.3391 0.2408 0 0.8106
DIVER2 7174 0.1730 0.1702 0 0.6571
Organizational Efficiency (OE) 4862 0.1049 0.1546 8.13E-05 1
Debt Margin (DM) 2737 0.4191 1.7534 4.82E-07 70.2558
Total Margin (TM) 4655 2.1369 10.2835 0.0021 583.4288
Size 4695 12.493 2.5014 4.5468 19.5310
Note. N refers to the total firm-year observations and is equal to 10,217 (i = 1,282; t = 8)
Appendix B
Summery statistics for the full sample, by Sector
Sector
Code
Growth
PE
Growth
TR
Growth
FA DIVER DIVER2 OE
Debt
Margin
Total
Margin Size
OW 0.5649 0.4094 1.7367 0.1870 10.0365 0.0806 0.3338 1.7651 11.5502
SH 0.0950 0.0542 2.3820 0.2633 0.1315 0.0772 0.3133 0.9949 13.0359
VU 0.1290 0.1123 0.1272 0.4593 0.2629 0.0352 0.3552 0.6616 14.8919
VO 0.3790 0.5508 3.9115 0.3951 0.2208 0.1244 0.3346 3.1043 13.3252
GE 1.5289 0.4362 1.3923 0.3220 0.1758 0.0969 0.3326 3.0822 12.2528
BL 0.4267 4.2549 3.6243 -0.0265 11.86865 0.1273 0.1273 4.7271 13.6474
MA 0.6986 0.2223 5.0650 0.1986 29.8043 0.1591 0.4077 3.0347 12.8241
ME 0.3700 0.1825 0.1902 0.3191 0.1603 0.0941 0.3594 0.8597 13.3072
CU 0.2244 0.6643 1.4014 -0.3571 166.7682 0.1210 0.4417 3.8623 14.5650
SR 0.2973 0.3240 0.2503 0.4674 0.2650 0.1274 0.4138 1.1319 13.7343
OO 1.2922 1.4590 13.0692 0.2704 0.1641 0.1048 0.6355 2.4280 11.3577
KL 0.0805 0.0600 0.4272 0.2920 0.1560 0.1153 0.2931 1.7718 13.3797
MI 0.0760 0.1464 0.3582 0.4407 0.2227 0.0959 0.3804 1.0541 13.5457
NA 0.1582 0.1074 0.1990 0.4603 0.2693 0.1065 0.3787 1.8771 14.9863
DI 0.1532 0.1421 0.4831 0.3635 1.1250 0.1037 1.3533 2.2001 12.7754
Total 0.5343 0.4026 2.4095 0.2372 15.0463 0.1054 0.4179 2.2635 12.5041
39
Appendix C
Correlation matrix
Growth
PE
Growth
TR
Growth
FA
Diver Diver2 OE Debt
Margin
Total
Margin
Size
Growth PE 1
Growth TR 0.0722* 1
Growth FA -0.0059 0.0007 1
Diver 0.0023 0.0060 0.0098 1
Diver2 -0.0037 -0.0073 -0.0004 0.9826* 1
OE -0.0239 -0.0121 0.0097 -0.0204 0.0139 1
Debt Margin -0.0042 -0.0063 -0.0144 0.0073 -0.0045 -0.0096 1
Total Margin -0.0100 -0.0196 0.0012 -0.1540* 0.1458* 0.1964* -0.0160 1
Size -0.0550* -0.0404* 0.0035 0.0228 -0.0014 -0.1260 -0.0600* 0.0494* 1
Note. The * denote statistical significance at the 5% level.
Appendix D
Correlation matrix
Growth
PE
Growth
TR
Growth
FA
Diver Diver2 OE Debt
Margin
Total
Margin
Size
Growth PE 1
Growth TR 0.0722* 1
Growth FA -0.0059 0.0007 1
Diver -0.0019 -0.0344* -0.0183 1
Diver2 -0.0032 -0.0226 0.0097 -0.9826* 1
OE -0.2710* 0.1257* -0.0070 -0.0204 0.0139 1
Debt Margin -0.0060 0.0110 -0.0085 0.0073 -0.0045 -0.0096 1
Total Margin 0.0282 0.0867* 0.0025 -0.1540* 0.1458* 0.1964* -0.0160 1
Size -0.0678 -0.1065* 0.0016 0.0227 -0.0014 0.1258* -0.0605* 0.0494* 1
Note. The * denote statistical significance at the 5% level.