The Perils of Free Cash Flow, Avoidance of Outside
Monitoring, and the Exploitation
of the Internal Capital Market
Jacqueline Garner1 and Adam Yore2
February 20, 2010
Abstract – All diversified firms are not created equal. Size matters when it comes to the functioning of corporate
internal capital markets and the value of corporate diversification. The “diversification discount” is largely driven by
mid-sized conglomerates with annual revenues of $20 to $400 million. These conglomerates experience an average
discount of 18% and significantly cross-subsidize their business units. By comparison, small and large
conglomerates suffer lesser discounts of 5% and 3%, respectively. This paper argues that the differences in excess
value are due to agency issues originating from managerial discretion over free cash flow and a lack of external
oversight. The results show that high levels of coincident free cash flow and corporate investment disrupt the
efficient operation of the firm’s internal capital market and lead to lower excess value. The free cash flow problem
appears concentrated in mid-sized and large conglomerates which produce substantially higher cash flow than small
conglomerates. Diversified firms that increase their disclosure or commit to pay out excess cash flow to shareholders
can substantially mitigate these negative effects. This study identifies which types of firms are most prone to the
dark side of corporate diversification. Institutional and activist investors should focus their monitoring efforts on the
mid-sized conglomerates which would otherwise go unwatched.
JEL classification: G31, G32, G34
Keywords: Corporate Diversification, Internal Capital Market, Free Cash Flow, SFAS 131
We are grateful to David Becher, Naveen Daniel, Eli Fich, Teresa Harrison, Natalia Reisel, Sam Tibbs, Ralph Walkling, the participants of the
2007 Financial Management Association’s and Southern Finance Association’s annual meetings, and to the seminar participants at Drexel
University for their helpful comments and suggestions. All errors and omissions are our own.
1 Department of Finance, Drexel University, Philadelphia, PA 19104
2 Department of Finance, Northern Illinois University, Dekalb, IL
The literature on corporate diversification has advanced under a few prevailing notions. First, the literature
contends that diversification destroys value and that this effect is unilateral across all firms. Contrary to this belief,
investors do not value all diversified firms equally. Size matters when it comes to the value of corporate
diversification and the functioning of internal capital markets. Second, recent studies suggest that the ability to
reallocate excess internally generated funds from one division to another gives the firm a real option to avoid costly
external financing [Yan (2006), Matsusaka and Nanda (2002)]. If diversified firms have excess free cash flow, their
ability to transfer capital across divisions is tantamount to an enterprise receiving capital without subjecting itself to
the approval and oversight of its investors. We examine both of these assertions in this paper.
There are economic rationales as for why the value of the diversified firm might vary by size. The most
commonly advanced explanation for the diversification discount is the inefficient use of resources by corporate
management, which leads to value destroying investment policies (Shin and Stulz, 1998; Scharfstein and Stein,
2000; Rajan, Servaes, and Zingales, 2000). Stulz (1990) shows that, rather than paying free cash flow to investors,
managers often prefer to reinvest it within the firm to maintain their corporate empire, even if the investments are in
negative net present value projects. It is generally recognized that the degree of financial constraints and level of free
cash flow varies by the scale and maturity of the firm’s operations. Fazzari, Hubbard, and Petersen (1988)
document significant differences between small and large firms in terms of financial constraints. Although they
show that nearly all firms rely extensively on retained funds, smaller firms have little cash flow after investment and
are more reliant on the external capital markets. Larger, more mature firms distribute greater fractions of their
internal cash flow, are less reliant on the capital markets, and spend relatively less on investment expenditures.
A diversified firm’s ability to avoid external capital markets is created by the capacity of the firm to
allocate capital across divisions (Matsusaka and Nanda, 2002). These arguments rely on the idea that multidivision
firms have a real option to finance their investments internally. However, since the option can be exercised at the
discretion of managers, it is not always done in the best interest of shareholders. That is, while competitive external
capital markets will allocate capital first to the highest q firms and avoid all firms with negative NPV investment
opportunities, as Jensen (1983) notes, management has the ability to allocate capital more socialistically or even to
value-destroying uses. Therefore, firms with high levels of free cash flow are better able to exploit this agency cost.
As discussed above, the extant literature has established a relation between levels of free cash flow and size (Fazarri,
Hubbard, and Petersen, 1988). For example, small firms typically have little excess free cash flow, in contrast to
their larger counterparts. Diversified firm CEOs with excess free cash flow are more likely to exploit the option that
Matsusaka and Nanda (2002) theorize for their own uses.
Jensen (1993) argues that outside control forces monitor management’s deployment of excess cash flow
and can possibly mitigate the agency issues associated with managerial discretion over excess capital. Consistent
with this assertion, Bens and Monahan (2004) show that thorough disclosure to the analyst community plays an
important role in monitoring investment behavior within conglomerates. If free cash flow, investment behavior, and
external monitoring differ by the size of the diversified firm as well as the levels of free cash flow it is plausible that
these varying firm characteristics will be associated with capital avoidance.
In this paper, we investigate whether these factors explain the cross-sectional differences in investment
efficiency and market valuation in conglomerates of varying sizes. In doing so, we present four contributions to the
literature. Our main contribution explores the characteristics, including size, that contribute to the ability of
diversified firms to avoid the external capital market. Diversified firms with high levels of free cash flow can fund
poor projects. While this agency cost can be short-circuited through external monitoring, high free cash flow firms
with inefficient capital markets are more likely to exploit the option to avoid outside capital markets.
Supporting our first contribution, we document that diversification discount varies with size. To our
knowledge, no study has thoroughly examined how the diversification discount varies with firm size and why this
relation might exist. Berger and Ofek (1995) present introductory evidence of how excess value from diversification
varies across size quartiles.1 However, their analysis does not fully explore the size-excess value relation, but instead
focuses on the existence of the discount itself. Using similar methodologies, this paper expands upon their initial
results by examining how the diversified firm’s value varies by size as well as why this relation exists.
The third contribution of this research is an examination of smaller conglomerates. The diversification
literature has excluded this important subset of the corporate universe by its focus on large, established
conglomerates and has almost completely omitted smaller diversified firms. This is due, in large part, to the success
of Berger and Ofek’s (1995) sample selection methodology which requires that sample firms produce at least $20
million in annual revenues. While academics and practitioners alike have learned much from their pioneering work
and the work of others that followed, we know almost nothing about the value impacts of diversification in smaller
1 See Berger and Ofek (1995) Table 4, Panel B.
firms or how these conglomerates manage their internal capital markets. The third aim of this study is to fill this gap
in the literature.
[Figure 1 and 2]
The results in this paper reveal economically and statistically significant differences in the excess value
attributed to diversified firms. As depicted in Figures 1 and 2, the empirical evidence indicates that mid-sized
conglomerates are valued significantly less than either small or large conglomerates and are the driving force behind
the observed “diversification discount.”2 The relation between firm size and the discount is persistent across
methodologies and subperiods. The data shows that mid-sized firms significantly cross-subsidize their business units
and operate inefficient internal capital markets. Further investigation reveals that these mid-sized firms possess
substantial levels of free cash flow, and the coincidence of excess cash flow and corporate investment is a significant
factor in explaining their lower firm values. Moreover, firms with inefficient capital markets and value-destroying
capital investment programs are more likely to avoid the external capital market, consistent with the argument that
managers engaging in empire building wish to avoid capital market oversight with Stulz, 1990. Indeed, the mid-size
firms are the worst offenders when we examine the interaction of free cash flow and investment on capital market
efficiency. Finally, the evidence suggests if these mid-sized conglomerates’ investment activities were subjected to
greater analyst scrutiny or if they committed to distribute free cash flow to shareholders, they would enjoy
substantially higher excess values thereby substantially reducing, or perhaps even eliminating, the diversification
The remainder of the paper is organized as follows. Section II presents a brief review of the conglomerate
literature. It also develops hypotheses for why internal capital market efficiency and excess value varies with free
cash flow and, consequently, firm size and how these characteristics are related to a firm’s ability to avoid the
monitoring of the external capital market. Section III discusses the sample selection and methodology and provides a
description of the observations. Section IV demonstrates how excess free cash flow negatively affects internal
capital market efficiency and firm excess value. Sections V and VI show how external control forces and the firm’s
payout policy moderates the agency issues associated with free cash flow. We conclude the paper in Section VII.
2 Figures 1 and 2 present evidence utilizing the universe of firms in the Compustat Business Segment database and only require that there is
complete data to compute both Berger and Ofek (1995) and Lang and Stulz’s (1994) excess value measures. All future tests employ the sample
restrictions documented in the Section III. As shown in Figure 3, the general conclusions are not sensitive to these restrictions.
Appendix A presents an algorithm for augmenting the SFAS 131 segment data so that it may be used in conjunction
with the SFAS 14 data.
II. Discussion and Hypotheses
Beginning with Lang and Stulz (1994) and Berger and Ofek (1995), the conglomerate literature has
repeatedly documented a “diversification discount.” That is, multi-segment firms suffer from valuation multiples
that are lower than those derived from a portfolio of their single-segment peers. One of the most common
explanations for the diversification discount is the misuse and misallocation of corporate resources within the
conglomerate firm. Existing evidence shows that management distributes capital socialistically within the
conglomerate firm rather than to their value maximizing uses. Berger and Ofek (1995) and Rajan, Servaes, and
Zingales (2000) show that the cross-subsidization of business units occurs and the subversion of corporate internal
capital markets destroys value.
Clearly the constrained manager would prefer to allocate scarce capital to its first best use, everything being
equal. There is little reason to believe that management would prefer to fund poor projects over good ones. Indeed,
some authors argue that the conglomerate structure can create substantial value for shareholders. Stein (1997) shows
that conglomerates can create value from corporate internal capital markets by engaging in “winner-picking” when
allocating capital among the firm's divisions. Matusaka and Nanda (2002) argue that efficient internal capital
markets can also add value by giving the conglomerate a real option to avoid the deadweight costs associated with
However, given ample slack, corporate headquarters has the ability to fund a greater universe of projects
and may overinvest rather than return excess capital to shareholders. Lamont (1997) provides evidence from the
1986 oil industry shocks where the subsidization of underperforming non-oil divisions was highly dependent on the
oil divisions’ free cash flow. Given this evidence, we hypothesize that those conglomerates which choose to invest
large amounts of free cash flow will do so in a manner that worsens the efficiency of their resource allocation
Large amounts of free cash flow may cause further problems with investment project selection, not just in a
relative sense, but in an absolute sense as well. Stulz (1990) and Jensen (1986, 1988) argue that management derives
private benefits from corporate investment and prefers to retain resources under its control. They show that
managers prefer to apply excess capital to potentially value-destroying uses rather than reducing their corporate
empire by distributing cash to shareholders. Lang, Stulz, and Walkling (1991) and Harford (1999) provide evidence
in support of this conjecture by showing that cash-rich firms tend to make value-destroying acquisitions. Doukas and
Kan (2004) extend this research to diversified acquisitions. They document that high free cash conglomerate bidders
conducting unrelated acquisitions experience significant declines in future profitability and excess value relative to
those conducting related acquisitions.
The product markets are a powerful controlling force over management (Jensen, 1993). Although they are
slow to act for large firms, it is likely that smaller firms are far more responsive to their discipline. Chevalier and
Scharfstein (1996) show that negative demand shocks significantly curtail investment expenditures for those firms
that are highly dependent on external capital. Further, Kedia (2006) and Karuna (2007) find that product market
competition does, in fact, alter managerial incentives. In small firms, low levels of capital are internally generated
capital, and raising additional external capital is often difficult. This demands that available resources are allocated
to their first-best use. Indeed, Vogt (1997) shows that value impacts of capital expenditure announcements are
decreasing in both free cash flow and firm size. We hypothesize that firms with high levels of free cash flow will
choose to invest it in negative net present value projects. Smaller conglomerates should be less prone to value-
destroying investment due to the scarcity of capital and dependence on the product markets.
Hypothesis 1a – Agency Costs of Free Cash Flow and the Resource Allocation Process:
With high levels of free cash flow, the unconstrained manager can socialistically fund investment projects
by choosing to invest both in positive and negative net present value projects. Consequently, the efficiency
of the resource allocation process within a corporate internal capital market should be decreasing with the
coincident levels of total firm investment and free cash flow. Analogously, the absence of resource
allocation process should be increasing with the coincident levels of total investment and free cash flow.
Since small firms are more closely aligned with both the product capital markets, the detrimental effects of
free cash flow should be significantly less for smaller firms than in larger firms.
Hypothesis 1b – Agency Costs of Free Cash Flow and Firm Value:
Firm excess value should be negatively related to the coincident levels of free cash flow and firm
investment due to management’s decision to fund too many investment projects as well as a sub-optimal
allocation of resources among the projects chosen. These effects should be limited to unconstrained larger
Outside controlling forces can be powerful motivators for management to utilize their investment resources
properly. Jensen and Meckling (1976) show that, if management can subject itself to effective monitoring, agency
costs may be reduced or eliminated. The monitoring forces can result from a number of sources including
information producers and providers of capital.
The recurrent need for capital presents an important avenue for managerial oversight. When raising capital,
managers essentially offer investors a referendum on their investment policy. Conglomerates renown for destroying
shareholder value may find the financial markets refusing their requests additional capital. If managers are
dependent on the financial markets for current and future infusions of capital, it is likely that they will curtail value-
destroying investments to maintain a reputation for value-maximizing behavior. Conversely, firms with the ability to
finance their investment with internally generated funds from other profitable divisions have a real option to avoid
this oversight for their capital expenditures. Consequently, these firms may choose to fund projects that would not
gain the approval of their shareholders or debtholders. This form of oversight should matter the most for those firms
who depend on the capital markets for recurrent financing. We hypothesize that the issuance of external capital will
improve resource allocation and firm excess value, especially in constrained firms.
Management often claims that the retention of excess cash flow to reinvest in the firm is essential for future
growth and the success of the firm. However, Easterbrook (1984) argues that management habitually overinvests
and advocates the use of dividends to mitigate this agency problem. Since dividends disgorge excess cash, the
constrained firm must often subject itself to the monitoring of the external capital markets each time it wishes to
make a major investment. Thus, by paying dividends managers commit themselves to future investor approval for
planned capital expenditures. The requisite approval of the firm’s investment program from the capital markets
serves to limit the free cash flow problem.
Empirical evidence suggests that the use of dividends for this purpose has the desired effect. Lang and
Litzenberger (1989) show that investors substantially reward overinvesting firms that commit to pay out free cash
flow in the form of dividends. The initiation or the increase of dividends represents a significant commitment as the
decision cannot be reversed without a substantial penalty to shareholder value.3 We hypothesize that a dividend
commitment will improve firm excess value for those firms with high levels of excess cash flow.
As a byproduct of their dealings with financial institutions when issuing capital, conglomerates often garner
a following in the analyst community. Further, management can encourage increased analyst coverage by improving
the quality of their disclosure (Lang and Lundholm, 1993; Francis, Hanna, and Philbrick, 1998). A number of
studies have found that, by uncovering and disseminating the value of managerial resource allocations, securities
analysts perform an important monitoring function for the firm’s shareholders (Moyer, Chatfield, and Sinsneros,
1989; Chung and Jo, 1996). Bens and Monahan (2004) argue that thorough disclosure to the analyst community
plays an important role in monitoring investment behavior within conglomerates. These authors also document a
positive association between disclosure quality and firm value. However, other information producers such as the
financial press, also perform a significant monitoring function (Dyck and Zingales, 2002) and might provide a viable
substitute for analyst oversight. Since Vega (2006) shows that media coverage substantially increases with the scale
of the firm, the benefits of analyst oversight are likely to be decreasing with conglomerate size. We hypothesize that
increased analyst coverage will improve resource allocations and firm excess value. Analyst oversight should be
most beneficial for those firms that are perceived to have an overinvestment problem and that lack other forms of
oversight from other information producers such as the financial press.
Hypothesis 2 – Monitoring:
We contend that firms with inefficient internal markets and high levels of free cash flow will avoid the
discipline of outside monitoring. However, if these firms are actually willing to subject themselves to
monitoring, the firm will benefit. Our monitoring hypotheses are as follows:
Firms that engage in value-destroying investments and/or cross-subsidize are less likely to issue capital and
dividends. These firms will also be associated with fewer analysts
3 Christie (1994) documents that dividend cuts and omissions are associated with 7% reduction in the firm’s stock price.
External monitoring pressure from information producers, the capital markets, and the market for
corporate control should mitigate the negative effect of free cash flow on the efficiency of the resource
allocation process and on overall firm investment. Proxies for the monitoring forces of the capital
markets should be positively related to excess value. Outside monitoring should be most beneficial for
those firms with high free cash flow and little oversight.
Given the above arguments, one expects that the conglomerates with the largest diversification discounts
should be those with the highest levels of coincident free cash flow and investment that also have the fewest analysts
following the firm, do not require external capital, , and have not made any commitment to disgorge excess cash
flow. One also expects that those firms who utilize their internal capital markets to pursue suboptimal investment
policies will choose to avoid the oversight of the external capital markets by refraining from issuing capital or
paying out dividends.
III. Sample Selection, Methodology, and Data Description
The initial sample consists of the universe of publicly traded multi-segment firms listed in the
COMPUSTAT industrial annual and business segment databases from 1978-2005.4 Firms began reporting financials
for their business segments in 1976 under FASB rule SFAS 14 and did so across industry lines. Material business
activities5 residing in a different 4-digit SIC code than the primary segment had to be reported as separate business
segments. FASB implemented SFAS 131 in 1998 which changed the manner in which firms report segment data.
Rather than reporting across industry lines, firms now must report business segments by internal organization. . It is
likely that the new data is superior to the old, since it is now representative of the firm’s actual resource allocations
rather than an artificial construct developed to conform to reporting standards. Rajan, Servaes, and Zingales (2000)
advocate the use of the SFAS 131 data due to its increased precision, but lament the fact that it was developed too
late to use within their study. However, the direct comparability of pre- and post-1998 segment data is compromised
4 We require one-year’s lagged data, so the final sample begins in 1979.
5 Operations are considered material if they possess sales, assets, or profits in excess of 10% of consolidated firm totals.
by the new FASB standard. It is thus necessary to augment the SFAS 131 segment data in order to use it in
conjunction with the SFAS 14 data. This process is described in the Appendix.
Each firm in the sample must have complete firm-level and segment-level data to compute all of the
explanatory variables used in the analysis.6 Following Berger and Ofek (1995), all financial firms (SIC codes 6000-
6999) as well as any firm with a financial business segment are excluded from the sample. The sum of segment sales
must also be within 1% of firm sales, and the sum of segment assets must be within 25% of firm assets. For those
firms where segment assets do not sum to firm assets, but do meet the 25% threshold, segment asset weights are
either grossed up or down by the percentage deviation between the sum of segment assets and firm assets to account
for the discrepancy wherever necessary.
In contrast to an overwhelming majority of the literature on conglomerates, we do not impose the Berger
and Ofek (1995) requirement that sample firms have at least $20 million in total sales. Including these otherwise
omitted firms allows the first glimpse of the impact/efficiency of the internal capital markets and the value
associated with diversification in small conglomerates. To help restrict the analysis to economically relevant firms
and to avoid valuation multiples with components close to zero (the rationale for the original size restriction in
Berger and Ofek, 1995), both firm sales and assets must exceed $1 million as measured in 1986 dollars. By
comparison, the smallest firm listed in the 1986 Standard and Poor’s (S&P) Composite had total assets of $1.24
million and net sales of $0.63 million.7
To maximize the separation in excess values among firm size constituencies, we partition the sample firms
into size groups with similar excess values according to the distribution in Figure 1. A natural separation in excess
value from the first decile to the second decile as well between the sixth and seventh decile appears. Firms are
classified as: 1) small, if they reside in the first decile of our sample in terms of deflated sales in 1986 dollars 2)
mid-sized, if they reside in the second through sixth deciles, and 3) large, if they reside in the seventh through tenth
deciles.8 The median sales for small firms are $10.22 million and range from $1.01 million to $19.0 million. Mid-
sized and large conglomerates have median sales of $130.9 million and $2,699.5 million and range from $19.0
million to $394.7 million and $395.0 million to $102,813.0 million, respectively.
6 This procedure is analogous to requiring firm-level data on market value of equity, book value of assets and equity, net sales, deferred taxes, tax
expense, interest expense, dividends paid, capital expenditures, investments, and acquisitions. This procedure also requires that each firm have
segment data on sales, assets, profit, and capital expenditures. Each of the firm’s segments must reside in an industry with at least five single-
segment firms so that we can minimize the noise associated with imputing Tobin’s q for the business segments.
7 The first year in the Berger and Ofek (1995) sample is 1986. For this reason, all values are deflated using the GDP Deflator (1986 = 100).
8 The results are similar if we utilize quintiles or deciles. For our sample, median sales for decile 1 are $10.2 million, median sales for deciles 2-6
range from $28.3 million to $ 294.7 million, and the median sales deciles 7-10 range from $520.1 million - $5,287.3 million.