Precautionary Savings with Risky Assets:
When Cash Is Not Cash
Ran Duchin, Thomas Gilbert, Jarrad Harford, and Christopher Hrdlicka*
July 2014
ABSRACT
We hand-collect data on the composition of corporate reserves, commonly called cash holdings, and find that about 40% of total reserves and 6% of total book assets are held in risky assets; moreover, about 80% of these assets are illiquid. While firms maintain safe reserves as precautionary buffers, risky reserves cannot be explained by the precautionary savings motive – they are concentrated in large, profitable firms, with “excess” reserves and low cash flow risk. We assess the optimality of this strategy and find that risky assets accentuate the taxation disadvantage of corporate investment income, implying that any gross value creation would have to be implausibly large. Indeed, we show that investors discount the value of a marginal dollar allocated to risky reserves. This activity represents a shadow asset management industry of about $1.5 trillion, with policy implications for disclosure and repatriation taxes.
JEL classification: G32, G34
Keywords: Liquidity, Cash, Investment Securities, Risk, SFAS 157
* All authors are at the Michael G. Foster School of Business at the University of Washington. Email: [email protected]; [email protected]; [email protected]; [email protected]. We thank Aaron Burt, Harvey Cheong, Matthew Denes, John Hackney, Lucas Perin, and especially Rory Ernst for excellent research assistance. We also thank seminar participants at City University-Hong Kong, Emory University, Erasmus University, the University of Amsterdam, the University of Hong Kong, the University of Kentucky, Tilburg University, Tulane University, Washington University in St. Louis, and the participants of the ASU Sonoran Winter Finance Conference 2014, the 9th Annual FIRS meetings, the 2014 SFS Cavalcade, the Pacific Northwest Finance Conference, the 2014 LBS Summer Finance Symposium, and the 2014 Western Finance Association Meetings for helpful comments.
2
The precautionary savings motive has been central to understanding corporate cash policy in
previous academic research. Starting with Keynes (1936), and extending through the models of
Baumol (1956), Miller and Orr (1968), and more recently, Kim et al. (1998) and Almeida et al.
(2013), most theoretical treatment begins with this primary objective of securing financing when
the firm may not have sufficient funds to invest or meet its obligations due to external finance
frictions. Indeed, this is the most common justification given by managers, as demonstrated by
the survey evidence in Lins et al. (2010) and Campello et al. (2011).
Empirically, researchers have had considerable success explaining cash holdings by
examining variation in firm characteristics tied to precautionary demand, such as cash flow
volatility, growth opportunities, and information asymmetry (see, for example Opler et al. (1999)
and Harford (1999)). Recent findings also support the importance of the precautionary savings
motive in explaining the dramatic increase in average cash holdings (e.g., Bates et al. (2009) and
Duchin (2010)) and underscore the importance of precautionary savings in mitigating the impact
of the 2008-9 financial crisis (e.g., Campello et al. (2010) and Duchin et al. (2011)).
A key assumption in these studies is that corporate “cash” reserves are in fact invested in
cash or highly liquid, risk-free near-cash securities, as would be necessary for them to form
precautionary savings. Recent anecdotal evidence in the press, however, suggests that corporate
treasuries have considerably broadened the scope of securities in which accumulated reserves are
invested. For example, the article “Google’s Latest Launch: Its Own Trading Floor”, published
in Business Week on May 27, 2010, reports that: “Google, it turns out, has launched a trading
floor to manage its $26.5 billion in cash and short-term investments… One of the company's
goals is to improve the returns on its money, which until now has been managed conservatively.”
3
In this paper, we investigate the composition of corporate reserves, traditionally called
cash holdings.1 We consider several nonmutually exclusive motives for investing corporate
reserves in potentially risky financial assets: 1) tax effects – the taxation of U.S. corporate
income and unrepatriated permanently reinvested earnings of foreign subsidiaries of U.S.
companies; 2) hedging benefits – financial assets that payoff in good states may finance growth
options in those states; 3) wealth transfers from bond holders via asset substitution – investing in
financial assets that earn the rate of return on the firm’s debt prevents windfalls to bondholders
resulting from additional safe reserves ; 4) earning alpha – corporate managers may consistently
earn positive abnormal returns on their investments in financial assets; and 5) agency conflicts
and overconfidence – corporate treasury personnel may be overconfident about their investment
skills, take excessive risks, and invest in financial assets to make their job more interesting or
develop their asset-management human capital, at the expense of the firm’s shareholders. We
discuss these hypotheses in detail in the next section.
Empirically, we present one of the first detailed analyses of the actual investments
making up corporate reserves2 with the aim of answering several research questions: What is the
composition of firms’ reserves and what fraction of corporate reserves is held in risky securities?
What are the characteristics of firms that take risk with their reserves? How do risky reserves co-
vary with the firm’s liquidity needs? Is there evidence of value creation?
1 Since we find that corporate reserves are often held in non-cash-like assets, we use the term “reserves” instead of “cash holdings” throughout the manuscript. 2 Brown (2012a and 2012b) uses the Federal Reserve Flow of Funds Accounts to document aggregate changes in financial asset holdings of non-financial firms. He argues that corporate market investments are getting riskier over time and are not a good store of cash. Cardella, Fairhurst, and Klasa (2014) analyze the split between cash & cash equivalents and short-term investments, arguing that firms are taking more risk with their short-term investments. We complement and extend these papers by hand-collecting the actual holdings at the firm level so that we can test cross-sectional hypotheses about the characteristics, determinants and consequences of firms’ holdings of risky assets.
4
Our empirical analysis exploits the introduction of the 2009 accounting standard SFAS
No. 157, which requires firms, for the first time, to report the composition and fair value of their
investment securities. Using hand-collected data from annual report notes, we undertake a firm-
by-firm analysis of the actual composition of reserves for industrial firms in the S&P 500 index.
Our evidence suggests that the types of investments vary widely and include domestic and
foreign corporate debt, foreign government debt, equity investments, mortgage and asset backed
securities, etc.3 These securities are clearly not risk-free, cash or near-cash securities. Hence, our
findings question the standard measure of cash holdings, defined as the Compustat variable Cash
and short-term investments (CHE), which, as we show, does include risky investment securities.
To illustrate this point, consider the example of Apple’s reserves as of the end of
September 2012. Apple held $121.2 billion in cash, short-term investments, and long-term
investments. Our analysis indicates that 76.1% of this amount was held in risky securities, which
included $46.8 billion in corporate securities (equities and bonds) and $12.0 billion in mortgage
and asset backed securities.
Apple’s example also illustrates that corporate reserves are often reported on the balance
sheet outside CHE. In Apple’s case, CHE does not include an amount of $92.1 billion held in
long-term marketable securities. Thus, CHE frequently underestimates the total value of
corporate reserves. This point is illustrated in Figure 1, which shows that reserves are often
included in other balance sheet items that comprise current assets.
We hand-collect complete information on corporate reserves, reconciling the balance
sheet with fair value data, and calculate new measures of total, risky, and safe reserves. We find
that on average, total reserves are 16.9% bigger than CHE, suggesting that the standard measure
3 More exotic examples include student loan backed auction rate securities, accounts receivable conduits, and Venezuelan and Greek bonds.
5
of cash holdings indeed underestimates actual corporate reserves. Thus, the recent build-up in
cash reserves documented by Kahle, Bates, and Stulz (2009) is even more pronounced when
considering firms’ total reserves.
In contrast to the conventional understanding of the role of reserves, our estimates
indicate that a surprisingly large fraction of reserves is held in risky and potentially illiquid
securities. Relative to the standard measure of corporate cash holdings (CHE), the average firm
in our sample held 25.5% of that value in risky securities. The magnitudes are even more striking
on a value-weighted basis, where the firms in our sample held a total of 47.8% in risky securities
relative to traditional measures of their cash holdings. Relative to total reserves, the average firm
in our sample held 16.6% in risky reserves on an equally-weighted basis and 38.3% on a value-
weighted basis. Overall, relative to its total book (market) value, the average firm in our sample
held 4.8% (3.8%) of its value in risky securities on an equally-weighted basis, and 5.8% (5.6%)
on a value-weighted basis. Risky reserves are also relatively illiquid, with approximately 57% of
these reserves reported as either moderately or very illiquid.
The risk of these assets comes from their poor performance at exactly the time firms
would need to draw on their precautionary savings. This risk is exemplified by the particularly
bad performance of many of these assets during the recent severe contraction in external
financing associated with the global financial crisis. These investments either lost considerable
permanent value or were held in illiquid assets that simply had no buyers for periods extending
to multiple years. Figure 2 demonstrates this point by analyzing the performance of investment
indices corresponding to the typical investment securities of the firms in our sample. As Figure 2
clearly shows, the typical corporate investment securities would have lost a substantial fraction
of their value during the crisis.
6
Next, we investigate which firms invest in risky securities. Consistent with extant studies,
we find that smaller firms, with more volatile cash flows and higher market-to-book ratios tend
to hold bigger overall reserves, as measured by both CHE and our measure of total reserves.
Prior studies view this evidence as supporting the precautionary savings motive. However, when
we separate between safe and risky reserves, we find that the precautionary savings motive only
explains safe reserves. In contrast, risky reserves are unrelated to cash flow risk, and tend to be
concentrated in larger, more profitable firms. These findings imply that a large fraction of
corporate reserves cannot be explained by the precautionary savings motive.
We also find that firms with more foreign income hold more risky assets as a fraction of
their total book assets. This finding is consistent with cash “trapped” abroad for repatriation tax
reasons (Foley et al. (2007)) being invested in risky securities. If this “trapped” cash is invested
in risky assets, then it might require an additional discount to that typically applied to these
reserves for their associated tax liabilities. Furthermore, since most investment securities are
domestic, our estimates indicate that from an economy-wide perspective, this cash is not
“trapped” abroad since firms are investing it in the U.S. financial markets.
While it is impossible to conclusively identify managers’ motivation in placing reserves
in risky securities, we provide some suggestive evidence. We first estimate an empirical model
of reserves and split firms into quintiles based on their excess reserve holdings. As Jensen (1986)
originally proposed, excess liquidity may exacerbate the firm’s agency costs. We find that firms
in the highest quintile of excess reserves also hold the greatest fraction of reserves in risky assets.
In particular, firms in the highest quintile hold, on average, 13.0% of their total assets in risky
securities, more than 5 times the risky holdings of firms in the lowest quintile (2.4%), and almost
three times the holdings of firms in the second-highest quintile (4.6%).
7
We also investigate whether proxies for the severity of the agency problem between
managers and shareholders are associated with the fraction of a firm’s reserves invested in risky
securities. We further include proxies for the managers’ overconfidence as well as managers’
stock- and option-based compensation to proxy for the incentive alignment between managers
and shareholders and for the incentive to take more risk and speculate. We find little evidence
that the proxies for the severity of the agency conflict between managers and shareholders are
correlated with risky investments. However, we do find that overconfidence as well as stock- and
option-based compensation are associated with investment in risky securities.
Our last set of analyses investigates the valuation of reserves invested in risky versus safe
assets. We estimate that the value of a marginal dollar invested in risky reserves is 12.9% to
21.5% lower than if it were invested in safe assets. We conclude that there is little support for
value creation. We note, however, that this speculative use of excess reserves may be better than
other alternatives. Investing in risky, but zero-NPV, financial assets rather than negative NPV
acquisitions (Jensen (1986) and Harford (1999)) is arguably better for shareholders. The tradeoff
depends on how often the suboptimal reserves investment strategy leads to underinvestment.
Our results have implications for the literature on cash holdings as they challenge the
precautionary savings explanation for an economically significant portion of corporate cash
holdings. They also highlight problems in using reported cash and short-term investments as a
measure of cash reserves in empirical studies. Our findings are also relevant to the debate over
whether and when excess reserves should be returned to shareholders through a change in payout
policy. In addition, they point to a further dimension on which lines of credit and cash reserves
differ for liquidity provision (see, for example, Sufi (2009), Disatnik et al. (2013), and Acharya
et al. (2013)) since lines of credit cannot be invested in risky assets.
8
Overall, our findings open new questions into the explanations for, and policy
implications of, what are essentially hedge funds operating within companies. Any investment
company managing more than $100 million of someone else’s money is heavily regulated and
faces disclosure requirements around its holdings and performance. In contrast, U.S. industrial
companies manage more than $1.5 trillion on behalf of their shareholders, with very minimal
disclosure requirements. Therefore, many shareholders cannot assess the investment strategy and
performance of this growing shadow asset management industry, leading to potential information
asymmetries between managers and investors.
The remainder of the paper is organized as follows. Section 1 discusses the theoretical
motives for, and implications of, holding risky reserves. Section 2 describes the data. Section 3
investigates the composition of corporate reserves. Section 4 studies the determinants and
implications of firms’ risky reserves. Section 5 concludes.
1. Theoretical Motivation
We consider several nonmutually exclusive motives for holding risky reserves: 1) tax effects;
2) hedging benefits; 3) wealth transfers from bond holders via asset substitution; 4) earning
alpha; and 5) agency conflicts and overconfidence. We discuss each in turn.
Taxes
We briefly discuss the tax effects here and provide a more formal treatment of taxes both in the
purely domestic case and when the reserves are held in foreign subsidiaries in Appendix C.
Under the current tax code, the double-taxation of U.S. corporate income creates a wedge
between the after-tax investment income a shareholder would receive investing on her own and
the after-tax investment income the shareholder would receive if a U.S. C-corporation invests on
9
her behalf. To fix ideas, consider a simple example. The C-corporation has $100 in reserves, the
tax rate on corporate income is 35%, the personal tax rate on dividend income is 20%, and the
expected return on the risky investment (dividend-paying equity in this case) is 10%.
If the corporation distributes the $100 as a dividend, the shareholder will have $80 to
invest at an expected return of 10%, or an after-tax expected return of 8%. Thus, at the end of
one year, the shareholder expects to have $86.40 after taxes. If the corporation retains and
invests the $100, it expects to earn $10 on the investment. The investment return flows through
comprehensive income and is taxed at 35%, so it expects to have $106.50 at the end of the year
to distribute to shareholders. After paying dividend taxes on the distribution, shareholders can
expect $106.50 x .80 = $85.20 after all taxes. This alpha of -$1.20 represents -1.5% of the $80
investment the shareholder could have made on her own. Thus, the corporation is starting from
an alpha of -1.5% and must be able to earn at least 1.5% alpha on its investments net of fees just
to bring its risky investments up to zero-NPV. It is worth noting that the magnitude of the tax-
induced negative alpha is increasing in the expected return (risk premium) of the investment.
Thus, the tax disadvantage of the corporation’s reserve investments is minimized in risk-free
assets.
Finally, it is worth mentioning that there are factors that may mitigate the detrimental
effect of corporate income taxation. Specifically, corporations are able to exclude 70% of
dividends received from investments in other U.S. corporations from taxable income (80% if
they own more than 20% of the other corporation). Note, however, that this exclusion only
applies to investments in U.S. corporate equity. There is no such tax break for investments in
corporate bonds. Interest income from municipal bonds is not taxable to anyone, including the
corporation. So, there are certain investment strategies (U.S. equities with high dividend yields
10
and municipal bonds) that would reduce the tax-induced negative alpha. Empirically, we will
show that reserve asset investments are not limited to these asset classes.
Hedging
In the presence of financing frictions, firms have a precautionary savings demand to hedge.
Depending on what the firm is trying to hedge, different implications arise. If the reserves are
held to mitigate the effects of aggregate cash flow shocks and varying costs of external finance,
holding any asset whose value positively covaries with aggregate cash flows is inconsistent with
the precautionary savings motive.
On the other hand, one can take the broader view that reserves are a means to move slack
from states where the firm does not need it to states where it does. In that case, a firm may have
more real options to exercise in aggregate good states and may value assets that payoff in those
states. Since real options are inherently hard to externally value and for insiders to credibly
communicate to external capital suppliers, internal financing of such options is less costly than
external financing. Thus, assets whose payoffs positively covary with the aggregate state of the
economy provide a poor hedge against aggregate adverse shocks, but may provide valuable
financing in good states of the economy in which the firm has real options to exercise.
In contrast, if the reserves are held to hedge against idiosyncratic, or firm-specific, cash
flow shocks, the covariance risk of the assets held in reserve is less important. The reason is that
as long as the payoffs of these assets are uncorrelated with the firm-specific shocks to cash flows
and investment opportunities, the firm’s ability to use its reserves is unaffected by the covariance
risk of its assets.
11
The implications reflect a potential balancing of hedging demands. On the one hand,
exposure to aggregate cash flow risk should lead firms to hold assets with lower covariance risk.
On the other hand, firms with pro-cyclical growth options would be expected to invest in assets
with covariance risk, balancing the need to provide some minimum slack in bad times with the
need to internally finance growth options in good times. We note, however, that some asset
classes, such as risky bonds, have asymmetric payoff functions, losing value in bad times
without having higher payoffs in good times. Holding such assets is therefore inconsistent with
the precautionary savings motive.
Asset Substitution
Firms typically issue risky, fixed-rate debt. Safer assets lower the required rate of return for
equity holders, but more importantly, raise the realized rate of return on the bonds because
default risk is lowered. Thus, the firm creates a windfall for bondholders by holding as cash
assets not needed to buffer cash flow shocks.
To avoid giving bondholders a windfall, the company should invest its excess reserves in
risky assets that earn at least the rate of return on their issued debt. Of course, the firm could
invest in riskier assets as well. However, in a repeated game, such substantial risk shifting would
have a costly impact on the firm’s reputation in the bond market. Hence, it is optimal for the firm
to invest in is assets that earn the same rate of return as its debt.
The implications of this strategy clearly depend on what the bondholders knew and
expected when the firm issued the bonds. It therefore applies most directly to cash accumulated
after the bonds were issued. Note that this strategy does not apply to all equity firms, and its
12
implications are limited for profitable firms with enough total reserves to be far from default
regardless of their investment strategy.
Alpha
If managers are able to earn excess risk-adjusted returns by investing the firm’s reserves in risky
assets, they are clearly creating value for the shareholders by undertaking positive NPV
investments in financial assets. The lack of disclosure, however, makes it impossible to assess
the performance of these investments directly. Nonetheless, there is a vast literature documenting
the absence of alpha in the money management industry (e.g., Carhart (1997)). Many of the
larger firms in our sample outsource the management of their reserves to the same pool of money
managers studied in this literature. For those that manage their money internally (such as Google
with its trading floor), it would be surprising to find that managers who can generate alpha are
hiding within non-financial companies (and not charging enough for their alpha-generating skills
to bring the net excess return to zero). Further, as we discuss above, the effect of taxes starts the
non-financial company in a negative alpha position, making it even more unlikely that
shareholders are receiving positive net alpha on these reserve assets.
A related argument is that there are scale efficiencies when the firm invests on behalf of
its shareholders. For example, the firm may be able to access certain private equity, hedge funds,
or other alternative investments that the individual investors cannot access on their own. We
note, however, that the majority of the typical firm’s shares are held by institutions, and this is
especially true of large firms with substantial reserves. Moreover, it is not clear what frictions
make it more efficient for a non-financial firm, rather than a financial institution, to act as an
intermediary on behalf of individual investors.
13
Agency Conflicts and Overconfidence
Another possibility is that risky reserves are simply another manifestation of the basic agency
problem between managers and shareholders. Typical characterizations of the agency conflict
focus on top managers, overinvestment and perquisites. In this case, the agency conflict is further
down in the organization where treasury personnel prefer to invest in securities other than
laddered U.S. Treasury portfolios, either to make their job more interesting or to develop human
capital that can be valuable elsewhere in the asset management industry. The latter is an example
of the conflict described in Holmstrom (1999) where an action’s returns to the manager’s human
capital are not positively correlated with the financial returns to investors. This creates an agency
conflict that is not mitigated by ex-post settling up in the labor market as described by Fama
(1980).
Even in the absence of agency conflicts, managers may still hurt shareholder value if
driven by confusion over the effect of low-yield investments on a firm’s ability to meet its cost
of capital.4 This is essentially the flip-side of the fallacy that debt is a cheap source of capital.
Under this motivation, management thinks it is acting in the interests of shareholders by reaching
for yield. Overconfidence on the part of managers could exacerbate this speculative, reaching-
for-yield motivation.
4 In discussing growing corporate cash reserves, one analyst remarked, “Corporations are flush with cash and that cash sitting in the corporate coffers is earning next to nothing. Companies have to do something with it." (Demos, T, Russolillo, S., and Jarzemsky, M. “Firms send record cash back to investors,” Wall Street Journal online March 7, 2013).
14
2. Data and Classification Methodology
In this section, we discuss our data collection and classification processes. We describe the
investment securities that firms hold and which make up their total financial reserves. We also
provide detailed examples of how firms report these reserves on their balance sheet and how this
reporting heterogeneity is taken into account in Compustat. Finally, we explain how we classify
these reserves as either safe or risky, liquid or illiquid.
2.1 Collection of Total Reserves and Investment Security Data
We construct firms’ total reserves by fully reconciling their balance sheets with hand-collected
data on their holdings of financial instruments from the fair value footnotes of the annual reports
(10-K) available on the Securities and Exchange Commission’s Edgar database. Since firms can
hold part of their reserves in other balance sheet accounts beyond the standard cash & cash
equivalents and short-term (marketable) securities, our extensive data collection exercise allows
us to both expand the standard Compustat definition of “cash” as given by data item CHE and
provide a detailed breakdown of the types of securities that firms hold as reserves.
To increase the transparency of the process used to calculate the fair value of their assets,
and primarily driven by the implementation of the Statement of Financial Accounting Standards
(SFAS) No. 157, most firms now report a footnote labeled “fair value measurements” in their
annual reports. This footnote typically includes the fair value of the firm’s financial holdings
broken down by asset class (bonds, equities, etc.) and valuation inputs (level 1, 2 or 3).5 SFAS
No. 157 requires firms to report the fair value of all financial/investment assets broken down by 5 In order to fully capture all fair value information of all reserves, including asset and level breakdowns, we regularly have to merge data reported across other footnotes such as “investments” and “accounting policies.” While SFAS No. 157 suggests that a table format for such reporting may be adequate, there is no formal requirement for the form of the reporting, and we therefore search for all the necessary information in both the text and the tables provided in the annual reports.
15
the type of inputs necessary to assess their fair value: quoted prices in active markets for
identical assets (level 1), significant other observable inputs (level 2), or significant unobservable
inputs (level 3). We use this level breakdown as a proxy for the liquidity of the reserves since an
increasing level signals an increasing difficulty for the firm to calculate the current fair value of a
given asset. More information about SFAS No. 157 and related accounting standards and rules
can be found in Appendix A.
In Panel A of Appendix B, we report a sample of the asset categories we collect in the
fair value footnotes and which are sorted by input level (level 1, 2 or 3).6 Even though firms’ fair
value footnotes often contain information about their liabilities and other assets that are not part
of their reserves, we do not collect data on derivatives (used for hedging or other purposes),
pension assets, deferred compensation assets, and assets held for strategic reasons (e.g., majority
shareholding in a subsidiary).7 We collect data on restricted cash even though its implicit (and
often regulatory-driven) illiquidity makes its inclusion as a reserve asset questionable.8
Excluding restricted cash from our analysis does not change any of our findings.
Our sample includes all firms that have been members of the S&P 500 Index at any point
between 2009, the year in which SFAS No. 157 became effective, and 2012.9 Following the
literature, we drop all financial and utility firms, which leaves us with a sample of 446 firms and
6 The same category can appear in multiple levels for the same firm and furthermore, different firms may report the same category as a different asset level. 7 More precisely, we exclude items labelled as “equity method investments” or “cost method investments,” as these mostly represent strategic investments, which are often not reported at fair value. We also ignore items labelled as “Rabbi trust assets” that are related to deferred executive compensation. 8 Compustat includes restricted cash in CHE if, and only if, the restricted cash account comes at the very top of the balance sheet, immediately after cash & cash equivalents and short-term investment securities. It is not included in CHE if the restricted cash account sits elsewhere on the balance sheet. 9 Any firm without an annual report during one (or more) fiscal year due to a merger or acquisition, for instance, is also dropped from the sample for that year. We also exclude the payroll processing firms ADP and Paychex since they behave as banks, holding large amounts of deposits on behalf of their customers.
16
1,727 firm-year observations spanning four fiscal years. We obtain monthly stock returns from
CRSP and firm-level accounting data from Compustat.
2.2 Mapping the Reserves and Linking Them to Compustat Items
We now explain how these data correspond to previous research on cash holdings and to the
standard data items in Compustat. Virtually all prior studies of cash holdings use the Compustat
item CHE, which is the sum of CH (cash) and IVST (short-term investments – total), to measure
the firm’s cash and cash equivalents. However, a significant portion of this “cash” need not be
held in safe and liquid assets, and moreover, firms may report additional security holdings
elsewhere on their balance sheets. As a result, we use the word reserves throughout the paper to
describe all financial assets (except for its derivatives, pension assets, deferred compensation
assets, and assets held for strategic reasons) that the firm holds.
Figure 1 illustrates these measurement issues using a diagram of a firm’s total reserves.
The firm’s total book value of assets (AT) comprises the standard cash and short-term
investments (CHE) in addition to various other assets (not represented). As noted above, both
components of CHE (i.e., CH and IVST) may include risky and illiquid assets over and above
the safe and liquid assets that the previous literature has assumed and focused on. Furthermore,
there can also be additional risky and illiquid assets reported in other accounts elsewhere on the
balance sheet.
While the Compustat data item IVAO (investments and advances – other) can include
some of the firm’s reserves, such as long-term investments, it can also include many other items
such as long-term receivables.10 ACOX (current assets – other) and AOX (assets – other), among
10 Compustat item IVAEQ (investments and advances – other) is about equity holdings in affiliates and subsidiaries where the firm has control. Since these are held for strategic reasons, they are not relevant for our analysis.
17
others, may also include some of the reserves as well as other items. As a result, since our goal is
to precisely measure the firm’s investment in risky securities, we cannot rely on Compustat data
and must collect this information from the fair value footnotes that provide a detailed breakdown
of the firm’s investment securities.
The firm’s total reserves, as shown by the grey boundary on Figure 1, encompasses all of
the firm’s cash & cash equivalents and short-term investments, as well as portions of other
accounts. Importantly, even though the fair value footnotes do not necessarily include all of the
reserves, we use the text surrounding the footnotes and financial statements to fully reconcile the
footnote information with the balance sheet accounts in order to get a complete representation of
the reserves
2.3 Examples of Security Holdings and Reporting
The reserves of companies include the safe and liquid assets that we expect them to hold for
precautionary savings: cash and cash equivalents such as Treasury bonds, money market funds,
commercial paper, and certificates of deposits. However, these reserves turn out to also include a
much wider range of assets that are both riskier and less liquid. We provide illustrative examples
of firms’ reserves, and the reporting of these reserves, for fiscal year 2012 in Appendix B.
The most cash-rich firms tend to have a particularly clear and thorough breakdown of
their financial assets, even though they are the exception rather than the norm. Google (Panel B
of Appendix B) has all of its reserves in CHE: $14,778 million in cash & cash equivalents and
$33,310 million in short-term marketable securities. However, as the tabulated fair value
footnote makes it clear, not all of its reserves are safe and liquid. For instance, it holds more than
$8 billion in mortgage- and asset-backed securities and over $2 billion in municipal securities.
18
Apple, based on their balance sheet, holds $10,746 million in cash and cash equivalents
(CH), $18,383 million in short-term marketable securities (IVST), and $92,122 million in long-
term marketable securities (IVAO). The total fair value of their financial instruments therefore is
$121,251 million, which is the number frequently quoted in the financial media.11 Even though
CHE is a severely under-estimated measure of Apple’s reserves, the tabulated fair value footnote
in Panel C of Appendix B provides a breakdown of all $121,251 million into asset classes and
level 1, 2, and 3. For instance, $3,109 million is reported as cash, $2,462 million is reported
under level 1 as mutual funds, and $20,108 million is reported under level 2 as U.S. Treasury
securities.
Panel D in Appendix B shows that, in the case of Intel, not all of their reserves are
reported in the fair value footnotes. Indeed, their cash & cash equivalent assets do not sum up to
the amount on the balance sheet (CH) since cash is missing from the footnote. Reconciling this
data with the balance sheet information is necessary in order to map out all of the firm’s reserves.
2.4 Asset Classification Methodology
For our sample of S&P 500 non-financial and non-utilities firms between 2009 and 2012, we
collect more than 2,500 separate holding types and their associated fair values from the footnote
of their annual reports and their balance sheets. With this data, our classification proceeds in two
separate steps. First, we classify each individual asset holding by type using the following set of
not-mutually-exclusive categories: cash, cash equivalents, money market fund, equity, debt,
corporate, government, agency, U.S., foreign, municipal, asset or mortgage-backed securities,
11 Compustat also reports aggregate values for the asset levels for some firms, but not all. For example, Apple’s AQPL1 (assets level 1 – quoted prices) reports $3,922 million, AOL2 (assets level 2 – observable) reports $114,370 million, and AUL3 (assets level 3 – unobservable) reports $0.
19
mutual fund, auction rate, time deposits, or commercial paper. For instance, a domestic equity
mutual fund would get classified into three categories: equity, U.S. and mutual fund.
Second, and most important for our analysis, we classify each asset holding as either safe
or risky. Almost all assets are clearly either in the safe category (e.g., cash, cash equivalents,
U.S. Treasury securities, time deposits, bank deposits, money market funds, and commercial
paper) or in the risky category (e.g., mutual funds, corporate bonds, mortgage-backed securities,
and foreign government bonds). However, some asset classifications are less clear: trading
securities, other investments, etc. In our classification, we operate under the assumption that a
firm has no incentive to hide the fact that it is holding cash if it is indeed holding cash. As a
result, if it chooses to name a portion of its holdings as “other”, we assume that it is not cash or
cash equivalent and hence that it is risky.
While the required reporting of asset levels (levels 1, 2, and 3) is a good proxy for the
liquidity of the securities, it is not informative about the riskiness of the securities since it only
pertains to the type of input required to assess fair value. Hence, a small cap equity mutual fund
would be a level 1 asset since its price is easy to obtain, even though it is clearly a risky holding.
In the Google and Apple examples presented in Panels B and C of Appendix B, the safe
assets are: cash, money market funds, U.S. Treasury securities, certificates of deposit and time
deposits, and commercial paper. We classify all other assets (e.g., mutual funds, marketable
equity securities, non-U.S. government securities, corporate securities, municipal securities, and
mortgage- and asset-backed securities) as risky. Level-1 assets are the most liquid, level-2 assets
are less liquid, and neither firm has any level-3 assets. As an example of a firm that does not
have a footnote detailing the breakdown of its asset holdings, FedEx specifies that the $4,917
million on its 2012 balance sheet is in cash and cash equivalents, which we label as safe.
20
3. The Composition of Corporate Reserves
3.1 The Asset Composition of Corporate Reserves
We begin our empirical analysis by presenting univariate results on the composition of
corporate reserves. Panels A and B of Table I report the breakdown of corporate reserves into
their primary asset classes, which comprise cash and cash equivalents, government debt,
corporate debt, asset-backed and mortgage-backed securities, other debt, equity, and other
securities. We further break down each primary asset class into the various securities that fall
under it, as well as domestic versus foreign securities. For each asset, Table I shows the fair
dollar value (column 1), and the fair value normalized by: book assets (column 2), market value
of equity (column 3), the standard measure of cash holdings, defined as cash and short term
investments on the balance sheet or Compustat’s CHE (column 4), and total reserves (column 5).
These ratios adjust for the scale of the company and its reserves, and facilitate the comparison of
our results with the traditional measures of cash holdings, which are typically normalized by firm
size. In addition, Table I also reports values for the aggregate categories: safe reserves, risky
reserves, and total reserves.
Panel A reports equally weighted averages calculated across all firm-year observations.
Based on panel A, the average firm invests a substantial portion of its reserves in risky assets. In
particular, the average firm invests about $1.2 billion in risky assets, which represents 4.8% of
total book assets and 3.8% of the market value of equity. Compared with Compustat’s CHE, the
standard measure of cash holdings, the average firm holds 25.5% in risky reserves. Based on our
comprehensive measure of reserves, which reconciles the firm’s balance sheet with its detailed
reporting of fair values, the average firm holds 16.6% of its reserves in risky assets.
21
Panel B reports aggregate values for 2012, the most recent year in our sample. Consistent
with Panel A, Panel B also indicates that firms hold a substantial percentage of their reserves in
risky assets. Specifically, the total value of risky reserves held by the firms in our sample in 2012
was $611 billion, collectively accounting for 5.8% of total book assets and 5.6% of the market
value of equity. Aggregated across all our sample firms, risky assets accounted for 47.8% of
CHE and 38.3% of total reserves. These estimates are very similar across the other years in our
sample (2009-2011), which are not reported for brevity.
The large estimates in Panels A and B suggest that firms hold a substantial fraction of
their reserves in risky assets that are not cash or near-cash, risk-free or near risk-free assets, and
therefore are not pure precautionary savings. Moreover, the differences between Panels A and B
suggest that larger firms hold more risky reserves - while the average firm in our sample holds
16.6% of its reserves in risky assets, the aggregate estimates are more than twice as large and
suggest that firms hold 38.3% of their reserves in risky assets on a value-weighted basis. This
result implies that while firms maintain safe reserves as precautionary savings, larger firms,
which are arguably less financially constrained and therefore have lower precautionary savings
demand, tend to invest more in risky, non-precautionary reserves.
More granularly, firms have substantial holdings of debt securities. Based on Panel A,
about 2% of firm value is invested in non-Treasury government debt, including municipal and
agency debt; about 1.5% is invested in corporate debt; and more than 0.5% is invested in other
debt securities, with the majority invested in asset-backed and mortgage-backed securities. Firms
also invest about 1% in other securities, including 0.3% in equity securities.
Further, the vast majority of corporate reserves are held in domestic securities. In fact,
foreign security holdings are negligible for all asset classes but government debt, where they
22
amount for 0.3% of total assets. To the extent that corporate reserves include substantial foreign
balances, arguably trapped overseas for tax considerations, as documented by Foley et al. (2007),
our estimates show that most of these reserves are invested domestically. These results highlight
a backdoor through which the money may flow back into the U.S. Thus, the seemingly-trapped
reserves are invested in the U.S. and therefore, from an economy-wide perspective, these
reserves are not trapped abroad.
In summary, a surprisingly large fraction of corporate reserves is held in non-cash
securities whose values co-vary with aggregate cash flows. The covariance risk of these assets
implies that they provide a poor hedge against aggregate cash flow shocks and may reflect
motives other than the precautionary savings motive. Our findings complement prior studies of
corporate cash reserves, such as Opler et al. (1999), Bates et al. (2009), Campello et al. (2010),
and Duchin et al. (2011), which focus on the precautionary savings motive and assume that
corporate reserves are invested in risk-free, cash or near-cash securities. In contrast, we offer a
detailed analysis of the non-precautionary portion of corporate reserves.
3.2 The Liquidity of Corporate Reserves
While covariance risk is a key dimension on which to evaluate corporate reserves,
liquidity is another important dimension of the precautionary savings motive. The assumption in
prior studies is that firms’ reserves comprise highly liquid assets (cash and cash equivalents) that
can be converted into cash at no or very low cost, thus providing a cost efficient hedge against
cash flow shocks when external finance is costly. In contrast, illiquid reserves impose a cost on
firms that wish to use their reserves, thus reducing the efficacy of these reserves as precautionary
buffers.
23
To measure liquidity, we collect data on the asset levels of corporate reserves. In
particular, SFAS No. 157 requires firms to report the type of inputs required to assess the fair
value of each reserve asset. Assets are grouped into three asset levels: Level 1 – quoted prices in
active markets for identical assets; Level 2 - significant other observable inputs; Level 3 -
significant unobservable inputs. Thus, level 1 assets are the most liquid, whereas level 3 assets
are the most illiquid.
Table II shows the fraction of each reserve asset class classified by the firm as level 1,
level 2, or level 3 assets. As in Table I, Panel A reports equally-weighted firm-level averages
over the entire sample period 2009-2012. Panel B reports aggregate sample-wide values as of
2012, the most recent year in our sample.
As expected, safe reserves are also highly liquid. The average firm reports 94.1% its total
safe reserves as level 1 (Panel A), and collectively, firms hold 86.1% of their safe reserves in
level 1 assets (Panel B). These findings are consistent with the role of safe reserves as
precautionary savings, since they are largely held in highly-liquid, cash or near cash securities
whose values do not co-vary with aggregate cash flows. Interestingly, however, the average firm
reports 5.9% of its safe reserves as illiquid (level 2 or level 3 assets), including 38.6% of
deposits, 84.0% of commercial paper, and 11.6% of money market funds. Collectively, firms
report 13.9% of their safe reserves as illiquid, as shown in Panel B. These findings show that
even among safe reserves, risk and liquidity are distinct attributes, with some safe reserves
invested in illiquid reserves that will likely impose a cost on firms attempting to use them as cash
buffers.
In contrast to safe reserves, risky reserves are also substantially illiquid. The average firm
holds 56.7% of its risky reserves in illiquid assets, with 74.0% of non U.S. Treasury government
24
debt classified as illiquid, 90.4% of corporate debt classified as illiquid, and 96.8% of asset- and
mortgage-backed securities classified as illiquid. In contrast, equity securities, which are
arguably riskier than debt securities, are largely classified as liquid assets (86.7% are classified
as level 1), once again highlighting the distinction between covariance risk and liquidity.
Panel B shows that collectively, on a value-weighted basis, the illiquidity of risky
reserves is even more striking. Compared to the average firm, which hold 56.7% of its risky
reserves in illiquid assets, firms collectively hold 78.8% of their risky reserves in illiquid assets.
These findings suggest that larger firms tend to hold more illiquid reserves. Coupled with the
results in Table I, our findings indicate that larger firms, with arguably lower precautionary
savings demand, tend to invest more in both risky and illiquid assets.
3.3 An Industry Analysis of Corporate Reserves
To investigate the characteristics of firms that take risk with their reserves, we start by
offering an industry analysis of corporate reserves. Panels A and B of Table III report the
holdings of risky reserves across the Fama-French 5-industry classification. Similar to Tables I
and II, Panel A reports firm-level averages, while Panel B reports aggregate numbers.
The estimates in Table III suggest that there are significant industry effects. Firms in the
Technology and Health sectors invest significantly more in risky assets than do firms in the other
sectors: The average Technology firm invests 11.1% of its total assets or 26.9% of its total
reserves in risky reserves and the average Health firm invests 7.4% of its total assets or 26.4% of
its total reserves in risky reserves. These values are substantially higher than those of firms in
other industries, which, on average, invest approximately 1.4-2.2% of their total assets and 10.1-
12.3% of their total reserves in risky reserves. As before, Panel B demonstrates that the
25
magnitudes are even more striking on an aggregate basis. Technology firms collectively hold
50.4% of their total reserves in risky reserves and Health firms hold 51.8% of their total reserves
in risky reserves. Our estimates suggest that contrary to the standard precautionary savings
motive, which implies that firms in growing, risky industries should hold cash and cash
equivalents, firms in the Technology and Health industries, which are characterized by volatile
cash flows and high growth opportunities, tend to invest in risky, non-cash reserves.
Panel C of Table III shows the top 20 firms that invest in risky reserves based on absolute
dollar amounts as well as fractions of total book assets and total reserves. Based on the dollar
amounts, the concentration of risky investment securities in the Technology and Health
industries is clear: out of the top 20 firms, 10 firms are in the Technology sector and 5 firms are
in the Health sector, with GE, Berkshire Hathaway, GM, Ford, and Coca-Cola being the
exceptions.
The estimates of risky reserves as a fraction of total assets further suggest that some firms
hold an extremely large percentage of their total assets in risky reserves. For example, Verisign
holds 70.6% of its total assets in risky reserves; Microsoft holds 57.3% of its total assets in risky
reserves; and Apple holds 52.4% of its total assets in risky reserves. Relative to total reserves,
the magnitudes of risky reserves are even more striking. Six firms out of the top 20 firms hold
more than 90% of their reserves in risky assets and 16 firms out of the 20 hold more than 80% of
their reserves in risky assets. The table is led by Microsoft, which, according to our estimates,
holds 97.1% of its reserves in risky assets.
In summary, our evidence indicates that risky reserves are concentrated in the
Technology and Health sectors. The assets of firms in these sectors are primarily intangible:
human capital and intellectual property. Pharmaceutical firms rely heavily on their drug patents
26
and Technology firms rely heavily on their patented electronic innovations. These firms do not
have significant tangible assets such as land, manufacturing plants (e.g., Apple outsources
manufacturing), etc., which could be pledged as collateral. Moreover, firms in these industries
operate in a volatile business environment, with high growth opportunities and many long-term
R&D investments. Thus, these firms have a strong precautionary savings motive, which cannot
explain their large holdings of risky reserves.
The concentration of risky reserves in growing sectors, however, may be consistent with
moving slack to aggregate good states in which firms have more growth options to exercise
(Carlson, Fisher, and Giammarino (2004)). Nevertheless, we are skeptical about the importance
of this motive since the results in Table I show that firms invest a large fraction of their risky
reserves in risky bonds, which have asymmetric payoff functions, losing value in bad times
without having higher payoffs in good times.
4. The Determinants and Implications of Corporate Reserves
4.1 Excess Reserves
We begin our analysis of the determinants of corporate risky reserves by investigating the
relation between risky reserves and large, or “excess”, reserve holdings. In particular, we test
whether firms with large, “excess” reserves tend to invest more in risky reserves. This line of
investigation is motivated by Jensen’s (1986) argument that “excess” reserves may push
managers to spend corporate resources inefficiently. Harford (1999), among others, provides
empirical support for this claim by showing that large cash holdings are correlated with
inefficient acquisition behavior.
27
In Panels A through C of Table IV, we sort firms into annual quintiles on three different
measures of corporate reserves, and calculate the average holdings of each reserve assets class,
normalized by the firm’s total book assets, across these quintiles. In Panel A, we form quintiles
based on Compustat’s CHE (cash and cash equivalents plus short term investments, as reported
on the firm’s balance sheet) normalized by total book assets, which is the traditional measure of
cash holdings used in the literature.
In Panel B, the quintiles are formed based on the traditional measure of “excess” cash,
again measured based on CHE divided by total book assets. More precisely, we estimate excess
cash holdings as the residual from the following regression model:
∙ ∙ ∙ ∙
where is firm i’s reported cash, cash equivalents and short-term investment holdings
(Compustat variable CHE), normalized by book assets, CF_VOL is the 10-year rolling window
volatility of cash flow/assets, MktToBook is the market-to-book ratio, CF is cash flow/assets,
Size is the natural logarithm of book assets, are indicators for the 5 Fama-French industries,
and are year dummies. Finally, in Panel C, we form quintiles on total reserves, our refined,
comprehensive measure of corporate reserves, which includes Compustat’s CHE, as well as
additional reserves reported elsewhere on the balance sheet.
The main message from Table IV is that risky investments are largely concentrated in
firms with the largest reserve and excess reserve holdings. Based on Panel A, which sorts firms
into quintiles on CHE, firms in the lowest quintile on traditional cash hold 0.6% of their total
book assets in risky reserves, whereas firms in the top quintile on traditional cash hold 15.5% of
their total assets in risky reserves. Furthermore, the increase in risky reserves from the lowest
quintile to the highest quintile is highly nonlinear, with firms in the second to fourth quintile
28
holding 0.82% to 4.9% of their assets in risky reserves. Thus, our results indicate that risky
reserves are largely concentrated in firms with the very largest cash holdings. In contrast, the
increase in safe reserves is relatively linear across the traditional cash quintiles, from 2.2% in the
lowest quintile, to 6.1%, 11.1%, 17.1%, and 27.0%, in the quintiles 2-5, respectively. These
findings further emphasize the concentration of risky reserves in firms with the very largest cash
holdings.
Moreover, the increase in risky reserves from the lowest to highest quintile is more than
twice as large as the increase in safe reserves. As a percent of total assets, the safe reserves of
firms in the highest quintile are 12.5 times those of firms in the lowest quintile. However, the
risky reserves of firms in the highest quintile are 26.3 times those of firms in the lowest quintile.
Thus, the increase in risky reserves at firms with the largest cash balances is disproportionally
high. Put differently, firms with the largest reserves invest a disproportionally high percent of
their total reserves in risky assets.
These effects hold across most individual categories of risky reserves. For example, firms
in the highest quintile hold 4.9% of their reserves in corporate debt, compared to 0.03% in the
lowest quintile. Similarly, asset- and mortgage-backed securities increase from 0.03% to 1.5% of
book assets, and non U.S. Treasury government debt increases from 0.1% to 6.9% of book
assets.
The results are similar if we sort firms on traditional excess cash or total reserves (Panels
B and C, respectively). For example, based on Panel B, firms in the top excess cash quintile
hold, on average, 13.0% of their book assets in risky reserves. In contrast, firms in the second
highest quintile only hold 4.6% of their assets in risky reserves, and firms in the lowest quintile
only hold 2.4% in risky reserves. The increase in risky reserves across the excess cash quintiles
29
is therefore also highly nonlinear. It is also disproportionally higher compared to the increase in
safe reserves: risky reserves increase by 5.5 from the lowest quintile to the highest one, whereas
safe reserves only increase by 3.7.
Finally, we note that in untabulated results, we find very similar results even after
excluding the top 20 firms on cash, excess cash, or total reserves. We therefore conclude that our
results are not driven by a small number of outliers.
Taken together, our results indicate a nonlinear relation between risky reserves and large,
“excess” reserves. Firms with the largest reserves invest 6-26 times more in risky assets than
firms with the smallest excess reserves. While the positive relation between large excess reserve
balances and investment in risky reserves may be interpreted as an inefficient, agency-driven
investment of the firm’s reserves, it is worth noting that the investment of corporate assets in
risky securities via the financial markets may be a less destructive form of agency costs than
empire building by CEOs through acquisitions or capital expenditures (e.g., Harford (1999)).
Given the presence of excess cash in the firm, shareholders and the board of directors would
rather see management invest this excess liquidity in efficient financial markets at fair prices
than attempt to spend it on potentially highly negative NPV projects. Investment in risky assets
could therefore be viewed as a lesser evil in terms of agency problems within the firm. Moreover
allowing management to invest "excess" liquidity in risky financial assets with impunity may be
an efficient outcome if shareholders and boards of directors cannot distinguish ex-ante good real
asset purchases from poor real asset purchases.
In Table V, we investigate the relation between liquidity and large, or “excess”, reserves.
As in Table IV, Panels A through C sort firms into quintiles on corporate reserves, measured by
standard cash holdings (CHE divided by assets), excess cash holdings, and total reserves,
30
respectively. The liquidity of corporate reserves is measured by asset levels, defined by SFAS
157 and reported by firms. We consider level 1 as liquid, and levels 2 and 3 as illiquid. In
addition to studying the relation between large reserves and the liquidity of total reserves, we
also investigate the relation between large reserves and the liquidity of risky reserves.
The results in Table V indicate that on average, firms with large, excess reserves hold a
larger percent of their reserves in illiquid assets. These results hold for all three measures of
corporate reserve holdings. Specifically, when we sort firms into standard cash quintiles (Panel
A), we find the firms in the lowest quintile hold 4.9% of their reserves in illiquid assets, whereas
firms in the highest quintile, with the largest cash balances, hold 35.1% of their reserves in
illiquid assets. Similarly, based on excess cash (Panel B) and total reserves (Panel C), firms in
the lowest quintile hold 9.3% and 5.8% of their reserves in illiquid assets, whereas firms in the
highest quintile hold 29.8% and 27.3% of their reserves in illiquid assets, respectively.
When we study the liquidity of risky reserves, we find a similar large increase in the
illiquidity of risky reserves across the reserve quintiles. As before, these findings hold for all
three measures of corporate reserves. For example, based on the standard measures of cash
holdings (Panel A) and excess cash holdings (Panel B), firms in the lowest quintile hold 40.7%
and 44.1% of their risky reserves in illiquid assets, whereas firms in the highest quintile hold
81.8% and 74.7% of their risky reserves in illiquid assets, respectively.
Collectively, our results show that the firms with the largest (“excess”) reserves invest a
substantially larger percent of their total and risky reserves in assets that are not only risky, but
also illiquid. These findings suggest a non-precautionary motive for the management of risky and
illiquid reserves, since the firms with the largest reserves hold reserves in excess of their
precautionary savings needs. Our results capture this idea indirectly by sorting firms based on
31
total cash holdings and reserves, as well as directly by sorting firms on excess cash, that is, on
cash in excess of the optimal cash reserves implied by the precautionary savings motive.
One interpretation of these findings is that in addition to managing precautionary buffers,
firms manage non-precautionary reserves for speculative motives. This behavior may be
consistent with Jensen’s (1986) argument that “excess” reserves may push managers to spend
corporate resources inefficiently. Under this interpretation, the firm’s investment in risky
reserves destroys shareholder value since it represents an agency conflict between shareholders
and managers, who choose to invest in risky reserves either to make their job more interesting, or
to develop human capital that can be valuable elsewhere in the asset management industry (e.g.,
Holmstrom (1999)).
Alternatively, both shareholders and managers may choose to speculate with their excess,
non-precautionary reserves in the financial market to “juice” their profits. Under this view,
shareholders and managers attempt to increase the value of their equity stake by speculating or
reaching for yield, possibly at the expense of the firm’s bondholders. Reaching for yield is also
consistent with an increased investment in illiquid assets as an attempt to capture the illiquidity
premium.
4.2 The Determinants of Risky Reserves
In this subsection, we conduct formal regression analysis of the determinants of the risk
and illiquidity of corporate reserves. Our baseline empirical model follows Opler et al. (1999)
and Bates et al. (2009), and includes explanatory variables that are motivated by the transaction
and precautionary motives for corporate reserves. Specifically, we include the following
variables:
32
(1) Market-to-book ratio – a proxy for the firm’s investment opportunities; (2) Size – to capture
the economies of scale in corporate reserves; (3) Cash flow – firms with higher cash flows might
accumulate more reserves and potentially invest more in risky reserves since they generate more
cash than required for precautionary reasons; (4) Net working capital – may include assets that
substitute for corporate reserves; (5) Capital expenditure –as shown by Riddick and Whited
(2009), a productivity shock that increases investment can lead firms to temporarily invest more
and save less, which would lead to a lower level of reserves; (6) Leverage – according to the
pecking order theory (Myers and Majluf (1984)), if debt is sufficiently constraining, firms will
use reserves to reduce leverage. In contrast, the hedging argument of Acharya, Almeida, and
Campello (2007) is consistent with a positive relation between leverage and reserve; (7) Cash
flow volatility – firms with greater cash flow risk hold more precautionary reserves. (8) Dividend
dummy – Firms that pay dividends are likely to be less risky and have greater access to capital
markets, so the precautionary motive is weaker for them; (9) R&D expenditure – firms with
greater R&D have higher growth opportunities and therefore greater costs of financial distress,
which increase their precautionary savings demand; (10) Acquisition expenditure – similar to
capital expenditure.
In addition to variables that capture the transaction and precautionary motives, our
augmented empirical model also includes foreign income to capture the repatriation tax-based
explanation of large reserve balances trapped abroad (Foley et al. (2009)). We would also like to
investigate the role of corporate governance and managerial incentives. To proxy for agency
conflicts between shareholders and managers, we use the G-index of minority shareholder rights
introduced by Gompers et al. (2003). To measure the incentive alignment between shareholders
33
and managers and their motivation to speculate, we use the CEO age as well as the fraction of
stock-based and option-based compensation of the top five managers in the firm.
We complement these measures with indirect proxies for managers’ motives to reach for
yield. Managers may be driven by confusion over the effect of low-yield investments on a firm’s
ability to meet its cost of capital. This is essentially the flip-side of the fallacy that debt is a cheap
source of capital. Under this motivation, management thinks it is acting in the interests of
shareholders by reaching for yield. To capture this motive, we include proxies for the firm’s cost
of capital: (1) the cost of equity –the CAPM beta, defined as the market beta and computed from
monthly returns, with the CRSP value-weighted index used as the market proxy; (2) the cost of
debt – measured by the firm’s credit rating.
Finally, overconfidence on the part of managers could exacerbate their speculative,
reaching-for-yield motivation. We follow Malmendier and Tate (2005) and measure CEO
overconfidence based on his stock option exercising behavior. Concretely, we define CEO
overconfidence as a binary variable equal to 1 if the CEO at some point during his tenure held an
option package until the last year before expiration, provided that the package was at least 40%
in the money entering its last year.
Table VI reports summary statistics for these explanatory variables. On average, the firms
in our sample have a market-to-book ratio of approximately 2, total cash flows that equal 9.3%
of book assets, leverage that equals 25% of book assets, and foreign income that equals 3.9% of
book assets. Thus, not surprisingly, the S&P 500 firms in our sample are highly profitable,
generate a substantial percent of their revenues overseas, and tend to be mature firms with
significant debt balances. Furthermore, our sample firms tend to pay dividends (our dividend
payer indicator equals 1 in 67.6% of the cases), and have an average beta of 1.2.
34
Table VI also provides summary statistics for the managers at our sample firms. Our
sample CEOs are 56.7 years old, on average, and are classified as overconfident according to
their option exercising behavior in 39.6% of the cases. The top 5 executive in our sample firms
receive 45.8% of their compensation in stock-based compensation and 27.3% of their
compensation in option-based compensation.
Table VII presents regression evidence explaining the risk of corporate reserves. Panel A
includes only the explanatory variables in the baseline regression model, whereas Panel B
includes the full set of explanatory variables discussed above. Each column in Table VII
corresponds to a different dependent variable. In column 1, the dependent variable is the standard
measure of corporate cash holdings, defined as Compustat’s CHE divided by book assets. We
include this regression model to facilitate the comparison of our results, which are based on
hand-collected data on corporate reserves, with the standard empirical model employed in the
literature. In column 2, the dependent variable is our comprehensive measure of corporate
reserves divided by book assets. In columns 3 and 4, we separate between safe and risky
reserves, both normalized by book assets. In column 5, the dependent variable is risky reserves
divided by total reserves. While the normalization of reserves by total book assets adjusts for
firm size, the fraction of risky reserves out of total reserves in column 5 captures the internal
composition of corporate reserves themselves. The regressions also include year and industry
fixed effects, to control for economy-wide shocks and industry-specific effects, respectively.
The baseline results in column 1 are consistent with prior empirical findings in the cash
literature (e.g., Opler et al. (1999)). Firms with greater demand for precautionary savings, as
proxied by their higher market-to-book ratios and cash flow volatility, tend to hold more cash.
Firms that pay dividends, as well as firms with higher leverage, capital expenditure, and
35
acquisition expenditure, tend to hold less cash. There are also economies of scale in cash
management, where larger firms hold less cash. When we re-estimate these regressions with our
comprehensive measure of corporate reserves, which reconciles the balance sheet and firms’ fair
value reporting, and includes reserves reported outside CHE on the balance sheet, we find very
similar results. These findings are reassuring since they suggest that our mismeasurement of total
corporate reserves using CHE did not lead to false inferences about the determinant of cash
holdings.
When we compare the determinants of safe and risky reserves in columns 3 and 4,
however, several novel findings emerge. First, while there are economies of scale in holding safe
reserves, larger firms do not hold lower risky reserves. In fact, based on column 5, the
composition of corporate reserves is tilted toward riskier reserves in larger firms. Second, the
positive relation between cash flow and corporate reserves only holds for risky reserves. In
contrast, firms do not increase their safe reserves when they generate higher cash flows. The
regression coefficient on cash flow in column 4 is insignificant in both Panels A and B, and has a
negative sign in Panel B. Further, based on column 5, firms hold a larger fraction of their
reserves in risky assets when they generate higher cash flows. This finding is consistent with the
notion that profitable firms, with a lower demand for precautionary savings, tend to speculate
more with their reserves. It also suggests that firms tend to park their free cash flows in risky
assets. We provide further evidence on the sources of risky reserves in section 4.4.
Third, our findings suggest that the positive relation between cash flow volatility and
corporate reserves is concentrated in risky reserves, whereas safe reserves are not significantly
related to cash flow risk. In fact, while not statistically significant, the results in column 5
36
indicate that cash flow risk is negatively related to the fraction of total reserves held in risky
assets.
In Panel B, we also investigate the relation between risky reserves and foreign income, as
a proxy for cash balances trapped abroad due to the repatriation tax. We find that both safe and
risky reserves, as a fraction of book assets, increase with foreign income. Thus, while firms
cannot distribute or spend this cash, they are able to invest it in the financial market. Our results
that they invest some of it in risky securities highlights a backdoor through which the money
may flow back into the U.S.; indeed, our analysis in Table I reveals that the firms’ investment
securities are mostly domestic. Thus, the seemingly-trapped cash is invested in the U.S. and
therefore, from an economy-wide perspective, this cash is not trapped abroad. We note, however,
that the composition of corporate reserves, as evident from column 5 of Panel B, is not
significantly related to foreign income.
We conclude this subsection with an investigation of corporate governance and
managerial incentives. We find little evidence that the G- index, which proxies for the severity of
the agency conflict between managers and shareholders, is correlated with risky reserves.
Consistent with Harford et al. (2008), however, we do find that firms with agency conflicts are
more likely to spend their safe reserves. While this non-result for the G-index may at first seem
surprising, we hypothesize that holding risky reserves may be viewed by shareholders and
directors as a lesser agency cost than reckless spending on large and permanent negative NPV
mergers and acquisitions, for instance. If markets are at least fairly efficient, then treasury offices
are buying these risky securities at fair prices and thereby earning the securities’ expected rates
of return. While these actions do expose the firm to a potentially large covariance risk, it may be
a less inefficient way of wasting the firm’s resources.
37
Consistent with a speculative motive, however, we do find that managerial
overconfidence and stock- and option-based compensation are associated with investment in
risky reserves. CEO overconfidence is positively related to risky reserves divided by book assets,
negatively related to safe reserves divided by assets, and positively related to the fraction of total
reserves invested in risky assets. Similarly, managers’ stock- and option-based compensation are
also positively related to risky reserves divided by either book assets or total reserves, and
negatively related to safe reserves divided by book assets. These effects are all statistically
significant at conventional significance levels and are economically meaningful. Based on
column 5, an overconfident CEO increases the fraction of total reserves held in risky assets by
2.7 percentage points. Further, an increase of one standard deviation in managers’ stock-based
(option-based) compensation corresponds to an increase of 1.2 (1.5) percentage points in the
fraction of total reserves invested in risky assets. These findings support the hypothesis that
shareholders and managers attempt to increase the value of their equity stake by speculating or
reaching for yield, possibly at the expense of the firm’s bondholders.
Finally, we note that all our results hold if we drop the top 10 firms, the top 20 firms, or
the top decile of firms in terms of cash holdings (Compustat’s CHE) or total reserves, i.e., our
results are not driven solely by Apple, Microsoft, and other extremely cash-rich firms.
4.3 The Determinants of Illiquid Reserves
In Table VIII, we present regression evidence explaining the liquidity of corporate
reserves. We include the same set of explanatory variables as in Table VII. Columns 1-3
correspond to the baseline regression model, whereas columns 4-6 include the full set of
explanatory variables. To study the liquidity of corporate reserves, we estimate regressions
38
explaining the ratio of illiquid reserves to book assets (columns 1 and 4), the ratio of liquid
reserves to book assets (columns 2 and 5), and the ratio of illiquid reserves to total reserves
(columns 3 and 6). The regressions also include year and industry fixed effects, to control for
economy-wide shocks and industry-specific effects, respectively. The standard errors are
clustered by firm.
The results in Table VIII show that the same set of firm-level characteristics that explain
holding risky reserves also explain holding illiquid reserves. More concretely, based on Table
VIII, larger firms hold more illiquid reserves, whereas smaller firms hold liquid reserves. In
addition, profitable firms that generate large cash flows also tend to hold more illiquid reserves.
We also find that cash flow risk explains liquid reserves, but is not significantly related to
illiquid reserves. Collectively, these findings suggest that firms with higher demand for
precautionary savings hold more liquid reserves, whereas firms with lower demand invest a
larger fraction of their reserves in illiquid reserves.
Based on the augmented regression model in columns 4-6, foreign income is positively
correlated with both liquid and illiquid reserves as a fraction of total book assets; it does not
explain, however, the holdings of illiquid reserves relative to liquid reserves. Similar to Table
VII, firms run by confident CEO hold more illiquid reserves and vice versa. We also find some
evidence that option based compensation is associated with a higher fraction of reserves invested
in illiquid assets.
4.4 The Sources of Risky Reserves
In Table IX, we present regressions designed to provide insight into where the investable
funds for safe and risky reserves come from. Specifically, Table IX follows the empirical model
39
in Kim and Weisbach (2006) and McLean (2011) and estimates panel regressions explaining the
sources of corporate reserves. The dependent variable is the natural logarithm of the annual
change in reserves, defined as follows:
1 . The explanatory
variables, which proxy for the potential sources of corporate reserves, include: (1) Cash flow -
net income plus amortization and depreciation; (2) Debt and equity issue - the cash proceeds
from debt sales and share issuance; (3) Other - all other cash sources, which include the sales of
property and the sale of investments. All 3 variables are scaled by lagged total assets. The
regressions include year and industry fixed effects, and the standard errors are clustered at the
firm level.
To facilitate the comparison of our results with previous findings, we consider in column
1 the sources of cash holdings measured by Compustat’s CHE. In columns 2-4, we consider our
measure of total corporate reserves (column 2), as well as total reserves broken down into safe
reserves (column 3) and risky reserves (column 4). Based on columns 1 and 2, the sources of
corporate reserves include both internally-generated cash flows and funds raised externally
through debt and equity issuance. Interestingly, when we use our modified measure of reserves,
we also find that other sources of funds, such as the sales of property and the sale of investments,
also contribute to the accumulation of corporate reserves.
More importantly, the main results in Table IX suggest that the sources for safe and risky
reserves are different. When firms raise outside capital and do not immediately use it to finance
real investments, that capital is more likely to be stored in safe assets. Conversely, free cash flow
is more likely to be invested in risky assets. This can be interpreted as further evidence in support
of an agency cost explanation. Easterbrook (1984) has pointed out that when firms raise external
capital, they submit themselves to monitoring and certification by capital providers and bankers.
40
This leaves them with less flexibility to invest precautionary savings in risky assets. Conversely,
free cash flow exacerbates agency conflicts in general, and, based on the results in column 4, free
cash flow not being put into real investments is often diverted into risky assets.
These results further suggest that outside investors do not provide firms with external
capital to fund investment in risky assets. Thus, our evidence is inconsistent with explanations
based on the ability of industrial firms to generate a positive alpha for investors, or run an
efficient investment fund that avoids the regulatory burden put on mutual funds and other
financial firms. If these explanations were indeed true, we would expect outside investors to fund
firms’ investment in risky assets.
4.5 The Value of Risky Reserves
Finally, in Table X, we examine the value effects of holding reserves in risky assets. To
do so, we employ the two most common approaches to estimating the value of a marginal dollar
of cash holdings, as developed in Fama and French (1998) (and utilized in Dittmar and Mahrt-
Smith, 2007), and Faulkender and Wang (2006).
Columns 1 and 2 present the Faulkender and Wang (2006) approach, based on the
following regression model:
, , ΔTotalreserves , ΔRiskyreserves , ΔE , ΔNA , ΔRD , ΔI ,
ΔD , Totalreserves , L , NF , Totalreserves , ΔTotalreserves ,
L , ΔTotalreserves , ,
where , , is a firm’s abnormal return, defined as the stock’s return during the fiscal year
less the return on the matching Fama and French (1993) size and book-to-market portfolio, ΔX
indicates a change in X from year t – 1 to t, Ei,t is earnings before extraordinary items, NAi,t is net
assets, RDi,t is R&D Expenses, set to zero if missing, Ii,t is interest expenses, Di,t is common
41
dividends, Li,t is leverage, defined as the book debt divided by book debt plus the market value of
equity, and NFi,t is net financing (including net equity issues and net debt issues). All variables
except leverage and excess return are deflated by the lagged market value of equity of the firm.
The regression in column 1 includes year fixed effects, whereas the regression in column 2
includes both year and firm fixed effects.
Following the critique in Gormley and Matsa (2014), columns 3-4 control for the
unobserved heterogeneity in returns by including annual dummies for the Fama and French
(1993) size and book-to-market portfolios. Therefore, in these columns, the dependent variable is
the unadjusted, raw stock return.
Because the dependent variable in columns 1-4 is a return and the variables are scaled by
the lagged market value of equity, the coefficients and can be interpreted as the value of a
marginal dollar invested in safe or risky reserves.
The point estimates in columns 1-4 suggest that while the marginal value of a dollar
invested in safe reserves is slightly over a dollar (ranging from $1.072 to $1.114), the value of a
dollar invested in risky reserves is 13.8 to 23.9 cents lower. These findings are consistent across
all the regressions in columns 1-4, and are statistically significant at conventional levels.
Columns 5-6, present the Fama and French (1998) approach, using the following
regression model:
, Totalreserves , Riskyreserves , E , dE , dE , RD , dRD ,
dRD , D , dD , dD , I , dI , dI , dNA ,
dNA , dMV , ,
where MVi,t is the market value of equity plus the book value of liabilities and dXt indicates a
change in X from time t – 2 to t. In columns 5-6, all variable are deflated by net assets. In the
Fama and French (1998) regressions, the variables of interest, and , capture the value of
safe reserves and the difference between the value of risky and safe reserves, respectively.
42
The inferences using the alternative approach in columns 5-6 are similar. The point
estimates suggest that the value of risky reserves is 23.2-29.7% lower than the value of safe
reserves. These findings hold after including firm fixed effects (column 6) and are statistically
significant at the 5% significance level.
Overall, the results from our analysis of the value of a marginal dollar invested in safe or
risky reserves continues to suggest an agency conflict over the stewardship of reserve assets.
Investors appear to recognize the downside of storing precautionary savings in assets with
covariance risk and value those reserves accordingly. It further suggests that a few vocal
investors and analysts notwithstanding, investors in general are not fooled into thinking that safe
assets are not earning their cost of capital and must be invested in risky or illiquid assets to
reach-for-yield.
5. Conclusion
This paper uses the introduction of a new accounting standard to offer some of the first evidence
on the investment securities that make up corporate reserves. We estimate that firms hold an
average of 17% (and a total of 38%) of their reserves in risky securities such as corporate and
foreign debt, and equity securities, whose payoffs co-vary with the firm’s operating cash flows.
These estimates question our standard measures of corporate cash reserves, which lump together
cash or near-cash securities and risky assets, and are inconsistent with the primary use of cash for
precautionary savings. This inconsistency is further emphasized by our findings that risky
investment securities are held by firms with high cash flow and low cash flow risk, whose
demand for precautionary savings is arguably lower.
We also find that investment in risky securities is highly concentrated in firms with
excess liquidity, whose reserves may be “trapped” due to repatriation tax considerations. These
43
findings suggest a form of an agency problem exacerbated by excess liquidity, which may push
managers to pursue private benefits at the expense of shareholders, or push both shareholders
and managers to speculate in an attempt to reach for yield, possibly at the expense of
bondholders. While we find little evidence pointing to a shareholder governance failure, we find
a positive relation between investment in risky securities and executive compensation tied to the
firm’s stock and stock options as well as managerial overconfidence.
Finally, we investigate the sources and value effects of risky versus safe reserve assets.
We find differences in both. Risky reserves are funded out of free cash flow while equity and
debt issues tend to go into safe reserves. Further, investors put a higher value on a marginal
dollar invested in safe reserves compared to one invested in risky reserves. We therefore
conclude that the most likely explanation for investing precautionary savings in risky assets is an
agency conflict combined with a mistaken desire to reach for yield on otherwise low-earning
reserve assets.
44
References Acharya, V., H. Almeida, and M. Campello, 2013, “Aggregate Risk and the Choice between
Cash and Lines of Credit,” Journal of Finance, forthcoming. Almeida, H., M. Campello, I. Cunha, and M. Weisbach, 2013, “Corporate Liquidity
Management: A Conceptual Framework and Survey,” Working paper: University of Illinois. Azar, J., J.-F. Kagy, and M. C. Schmalz, 2014, “Can changes in the cost of cash resolve the
corporate cash puzzle?,” Working paper: University of Michigan. Bates, T., K. Kahle, and R. Stulz, 2009, “Why Do U.S. Firms Hold so Much More Cash than
They Used to?” Journal of Finance, Vol. 64, pp. 1985-2021. Baumol, W. J., 1952, “The Transactions Demand for Cash: An Inventory Theoretic Approach,
Quarterly Journal of Economics, Vol. 66, pp. 545–556. Bebchuk, L., A. Cohen, and A. Ferrell, 2009, “What Matters in Corporate Governance?” Review
of Financial Studies, Vol. 22, pp. 783-827. Brown, C., 2012, “Corporate Market Investments: An Examination of the Cash Storage View”,
Working paper: National University of Singapore. Brown, C., 2012, “Corporate Market Investments: Risk, Returns and Governance”, Working
paper: National University of Singapore. Campello, M., E. Giambona, J. R. Graham, and C. R. Harvey, 2011, “Liquidity Management and
Corporate Investment During a Financial Crisis,” Review of Financial Studies, Vol. 24, pp. 1944-1979.
Campello, M., J. R. Graham, and C. R. Harvey, 2010, “The Real Effects of Financial
Constraints: Evidence from a Financial Crisis,” Journal of Financial Economics, Vol. 97, pp. 470–487.
Cardella, L., D. Fairhurst, and S. Klasa, 2014, “What determines the composition of a firm’s
total cash reserves?,” Working paper: Texas Tech University. Carlson, M., A. Fisher, and R. Giammarino, 2004, “Corporate Investment and Asset Price
Dynamics: Implications for the Cross-Section of Returns,” Journal of Finance, Vol. 59, pp. 2577-2603.
Disatnik, D., R. Duchin, and B. Schmidt, 2013, “Cash Flow Hedging and Liquidity Choices,”
Review of Finance, forthcoming. Dittmar, A. and J. Mahrt-Smith, 2007, “Corporate Governance and the Value of Cash Holdings,”
Journal of Financial Economics, Vol. 83, pp. 599-634.
45
Duchin, R., 2010, “Cash Holdings and Corporate Diversification,” Journal of Finance, Vol. 65, pp. 955–992.
Duchin, R., O. Ozbas, and B. A. Sensoy, 2010, “Costly External Finance, Corporate Investment,
and the Subprime Mortgage Credit Crisis,” Journal of Financial Economics, Vol. 97, pp. 418–435.
Fama, E., 1980, “Agency Problems and the Theory of the Firm, Journal of Political Economy, Vol. 88, pp. 288–307. Foley, F., J. Hartzell, S. Titman, and G. Twite, 2007, “Why Do Firms Hold So Much Cash? A
Tax-Based Explanation,” Journal of Financial Economics, Vol. 86, pp. 579–607. Gompers, P., J. Ishii, and A. Metrick, 2003, “Corporate Governance and Equity Prices,”
Quarterly Journal of Economics, Vol. 118, pp. 107–155. Gormley, T. and D. Matsa, 2014, “Common Errors: How to (and Not to) Control for Unobserved
Heterogeneity,” Review of Financial Studies, Vol. 27, pp. 617–661. Harford, J., 1999, “Corporate Cash Reserves and Acquisitions,” Journal of Finance, Vol. 54, pp.
1969–1997. Harford, J., S. Mansi and W. Maxwell, 2008, “Corporate Governance and Firm Cash Holdings,”
Journal of Financial Economics, Vol. 87, pp. 535-555. Holmstrom, B., 1999, “Managerial Incentive Problems: A Dynamic Perspective,” The Review of
Economic Studies, Vol. 66, pp. 169-182. Jensen, M., 1986, “Agency Costs of the Free Cash Flow, Corporate Finance and Takeovers,”
American Economic Review, Vol. 76, pp. 323–329. Jensen, M. and W. Meckling, 1976, “Theory of the Firm: Managerial Behavior, Agency Costs
and Ownership Structure,” Journal of Financial Economics, Vol. 3, pp. 305–360. Keynes, J.M., 1936, “The General Theory of Employment, Interest, and Money,” Harcourt
Brace, London. Kim, C., D. Mauer, and A. Sherman, 1998, "The Determinants of Corporate Liquidity: Theory
and Evidence," Journal of Financial and Quantitative Analysis, Vol. 33, pp.335–359. Kim, W. and M. Weisbach, 2008, “Motivations for Public Share Offers: An International
Perspective,” Journal of Financial Economics, Vol. 87, pp. 281–307. Krueger, P., A. Landier, and D. Thesmar, 2011, “The WACC fallacy: The real effects of using a
unique discount rate,” Working paper: University of Geneva.
46
Lins, K., H. Servaes, and P. Tufano, 2010, “What Drives Corporate Liquidity? An International Survey of Cash Holdings and Lines of Credit,” Journal of Financial Economics, Vol. 98, pp.160–76.
Malmendier, U. and G. Tate, 2005, ”CEO Overconfidence and Corporate Investment,” Journal
of Finance, Vol. 60, pp. 2661-2700. McLean, D., 2011, “Share Issuance and Cash Savings,” Journal of Financial Economics, Vol.
99, pp. 693–715. Milbradt, K., 2012, “Level 3 Assets: Booking Profits and Concealing Losses,” Review of
Financial Studies, Vol. 25, pp. 55-95. Miller, M. and D. Orr, 1966, “A Model of the Demand for Money by Firms,” Quarterly Journal
of Economics, Vol. 80, pp. 413–435. Opler, T., L. Pinkowitz, R. Stulz and R. Williamson, 1999, “The Determinants and Implications
of Cash Holdings,” Journal of Financial Economics, Vol. 52, pp. 3-46. Sufi, A., 2009, “Bank Lines of Credit in Corporate Finance: An Empirical Analysis,” Review of
Financial Studies, Vol. 22, pp. 1057–1088.
47
Appendix A: SFAS No. 157
The Statement of Financial Accounting Standards No. 157 entitled “Fair Value Measurements”
requires corporations to disclose the process by which they obtain the fair value of all the
financial assets held on their balance sheet. More precisely, SFAS No. 157 has three main
objectives: defining fair value, establishing a framework for measuring fair value, and expanding
disclosure requirements about fair value measurements, all within the generally accepted
accounting principles (GAAP). As stated, “fair value is the price that would be received to sell an
asset or paid to transfer a liability in an orderly transaction between market participants at the
measurement date.” If a market price for an asset is not easily available in the market, then the
fair value must be estimated and the valuation assumptions must be disclosed in a transparent
way.
Based on the availability and reliability of a market price, and the potential assumptions
and inputs needed to estimate a price, every asset falls into an asset level hierarchy that is divided
into three levels (1, 2, and 3). A level 1 asset is an asset for which a reliable market price is easily
available and no other inputs are required to assess fair value. Two examples are cash and large-
cap U.S. equity mutual funds. Such assets are typically highly liquid instruments traded on an
exchange. A level 2 asset is an asset for which the assessment of fair value requires another
observable input besides an easily available price. An example is an interest rate swap based on a
specific bank’s prime rate. A level 3 asset is an asset for which unobservable inputs are required
in order to assess fair value. If there is no market price for a given asset-backed security, for
instance, then a valuation model must be used, requiring a number of assumptions. These inputs
must be disclosed along with the estimated asset value.
48
SFAS No. 157 was first issued in September 2006 and was first implemented for financial
statements issued for fiscal years starting after November 15, 2007. After a year of transition, it
became fully effective for fiscal year 2009. While it is obviously not the first statement about fair
value measurements (others are SFAS No. 107, 133, and 155), SFAS No. 157 greatly increases
the disclosure requirements and puts a lot of emphasis on market-specific measurement and not
firm-specific measurement. This therefore forces corporations to disclose a clear breakdown of
their assets based on the assumptions they make when assessing fair value.
Appendix B: Examples of Asset Categories and of Fair Value Tables
Panel A: Asset Classification into Levels 1, 2, or 3 The panel shows a sample of the asset categories found in the footnotes of annual reports that disclose and explain the fair value of the assets held by firms. Following the guidelines in SFAS No. 157, firms classify their investment holdings into level 1, 2, or 3 depending on the input(s) required to determine their fair value. Level 1 Level 2 Level 3 Cash Cash equivalents Mutual funds U.S. Treasury securities Equity securities Corporate bonds – non U.S. Available-for-sale securities Bank deposits Money market funds
U.S. Treasury Securities Commercial paper Corporate bonds Time deposits Corporate bonds – non U.S. Asset-backed securities Available-for-sale securities U.S. Agency securities Government bonds – non U.S. Municipal bonds and notes Other securities
Venture capital investments Corporate bonds – non U.S. Available-for-sale securities Closed-end municipal bond funds Mortgage-backed securities Auction rate securities Long-term Venezuelan bonds Convertible debt securities
49
Panel B: Google (GOOG) Google’s reserves are all contained in CHE, where CH is the $14,778 million in cash and cash equivalents, and IVST is the $33,310 million in marketable securities.
Panel C: Apple (AAPL) Beyond CHE, Apple has an additional $92,122 million in reserves under long-term marketable securities (IVAO in Compustat).
50
Panel D: Intel (INTC) Intel’s fair value footnote includes its liabilities as well as some derivative assets and loans receivable, which we exclude. However, it is missing parts of its cash & cash equivalents, namely the cash, which we reconcile using the balance sheet. The footnote only tabulates $7,885 million of cash equivalents, which implies that the firm has $593 million in cash in order to sum up to the $8,478 million of cash & cash equivalents on the balance sheet.
51
Appendix C: Tax Effects
Formally, one can show that in general, when the corporate tax rate is greater than the
individual investment tax rate, the present value of this loss is increasing in both the expected
investment return and investment horizon. To see this result, let be the corporate distribution
tax rate (e.g., dividend tax rate), be the corporate earnings tax rate and be the individual
investment earnings tax rate (e.g., capital gains rate).
Consider a financial investment with its only gain realized at a horizon of years with
annual expected gross return . The future value of each dollar within the firm that is
immediately distributed to the individual investor and invested by the individual in this project is
1 1 1 .
The first term is the gross investment return and the second term is the taxes paid on the net
return. The future value of each dollar that the firm instead invests itself for years before
distributing is
1 1 .
The difference in future value between the firm investing and the individual investing is
1 1 ,
which is negative whenever the corporate tax rate is higher than the individual investment tax
rate.
If we assume that an individual facing marginal tax rate is the marginal investor then
the appropriate compounded risk adjusted discount rate is
1 1 1 .
Using this discount rate, the present value of the difference between the firm investing and the
individual investing is
52
1 11 1 1
.
Again, if the corporate tax rate is greater than the individual investment tax rate, there is a loss in
present value when the firm invests rather than the individual. This loss is increasing in both the
investment horizon and the riskiness (return) of the investment undertaken, so it is minimized by
investing in low return, low risk assets, i.e., the risk free asset. The limit of this present value as
the riskiness or investment horizon goes to infinity is
11
.
When the corporate tax rate is higher than the individual rate, the absolute value is the maximum
loss the firm can create by investing in risky assets. With the tax rates from thr preceding
numerical example, this maximum loss is 19% of the value that an investor would receive from
an immediate distribution.
If the taxability of investment gains cannot be delayed to the terminal period, e.g., the
firm invests in interest bearing securities, the losses from the firm investing on investors’ behalf
are even larger. For example, if investment gains accrue and are taxed yearly (e.g., interest
payments), investing in risky assets with a 10% rate of return yields a 1.4% loss in present value
(e.g., negative alpha) every year.
Next, we consider reserves held by foreign subsidiaries. If the firm can delay investment
gains until the end of the years, there is a tax benefit of investing in risky assets when the
foreign corporate tax is less than the U.S. individual investment tax rate ( , and a tax
disadvantage otherwise. However if the firm cannot delay these investment earnings, then the
foreign tax rate must be strictly smaller than the individual tax rate for a tax benefit. This
breakeven foreign tax rate is decreasing in the investment horizon and expected rate of return;
53
for a large enough value of either, there is no positive foreign tax rate that creates a tax
advantage for the firm to delay repatriation to invest in financial assets.
To understand the consequences of investing foreign reserves, one must distinguish
actual tax payments from the accounting recognition of these payments. Taxes on the earnings
that are the source of these reserves are paid in the foreign jurisdiction at the time the revenue is
earned. The difference between those taxes paid and the typically-higher U.S. tax rate is owed
upon repatriation.12 Reserves earned and held in foreign subsidiaries of U.S. companies generally
have the accounting designation of “permanently reinvested earnings (PRE). This designation
does not alter the timing of the tax payments. Such designation only alters the firm’s accounting
recognition of the additional U.S. taxes owed. Thus, when firms designate foreign reserves as
PRE, they carry them on their books at their pre-repatriation tax amount, ignoring the eventual
tax liability due.
The investment of foreign reserves has a separate tax treatment from that of the principal
and is unrelated to the designation of the principal as PRE. Foreign reserves can be invested in
any kind of securities, both foreign and U.S. Importantly, realized earnings on such investments
are almost always considered passive and are immediately taxable, just as investment earnings
on U.S.-based reserves are, at the level of the U.S. parent as “Subpart F” income, regardless of
the source of the investment.13 We extensively discussed tax issues surrounding cash held
abroad with a global tax partner at a big-four accounting firm and the bottom line of our
discussions is that there does not appear to be a tax arbitrage on investment earnings by investing
12 If repatriation occurs in a period when a firm has an operating loss, this loss may be used to offset the repatriated foreign earnings. 13 Firms can avoid investment income being classified as Subpart F income if their foreign subsidiary is not classified as a controlled foreign corporation (CFC), by having less than a 50% stake in the subsidiary. Even if classified as a CFC, the investment income can avoid being classified as Subpart F income if (1) the CFC resides in a jurisdiction with a tax rate at least 90% that of the U.S., (2) the investment income is less than $1 million per year, or (3) the investment income is less than 5% of the total income of the CFC that year.
54
in risky assets with foreign cash.14 However, there can be a tax benefit of delaying the taxation
of the investment principal if the foreign corporate tax rate is low enough and the firm can delay
investment earnings until the end of the investment horizon. In such a case, this benefit increases
in the horizon and expected return of the investment. This benefit becomes a cost if the foreign
taxes are higher than the individual rate.
To see this potential benefit (cost), we continue with the same notation and assumptions
as in the previous calculations. But now also let be the foreign corporate tax rate. The future
value of the firm immediately repatriating and distributing a pre-tax dollar of foreign earnings is
1 1 1 .
And the future value of the firm delaying repatriation and investing the reserves itself is
1 1 1 .
The difference in future value between the firm investing and the individual investing is
1 1 1 .
This difference is positive only when the foreign corporate tax rate is less than the U.S.
individual investment tax rate.
If, again, we make the assumption that the marginal investor is the individual, we can
calculate the present value of this difference:
1 1 11 1 1
.
This present value is increasing in magnitude in both the expected rate of return and investment
horizon. The present value is positive when the foreign corporate tax rate is lower than the
individual U.S. investment tax rate. Taking the limit as the horizon or expected return goes to
infinity we see the maximum (minimum) of the present value is
14 For further information on the taxation of foreign reserves see chapter 10 of Scholes et al. (2009).
55
1 11
.
When the firm, however, is unable to invest in assets that have long delayed capital gains,
the advantages are reduced, which is important because the firms in our sample are heavily
investing in risky debt rather than equity. For the firm’s investment of its returns to be beneficial,
the foreign tax rate must be strictly less than the U.S. individual tax rate for any horizon over one
year. The breakeven foreign tax rate is a decreasing function of the investment’s horizon and
expected return. And for sufficiently high expected returns or long enough investment horizons,
there are no positive foreign tax rates that make delayed repatriation and disbursement for the
purposes of financial investment a superior option.
56
FIGURE 1 Breakdown of Corporate Reserves
This figure shows a hypothetical breakdown of a firm’s financial reserves (not to scale and purely for illustrative purposes). Out of the total book value of assets (AT), a certain percentage is held as cash & cash equivalents and short-term investments (CHE = CH + IVST). A certain percentage of CHE may be invested in risky and/or illiquid assets. In addition, the firm may hold more risky and/or illiquid assets elsewhere on its balance sheet, for instance under long-term investments or other assets, some of which may be listed under IVAO in Compustat.
57
FIGURE 2 The Performance of Investment Indices during the Recent Financial Crisis
This figure reports the performance of nine investment indices that correspond to different asset classes from August 2007 to July 2009. The levels of all indices are normalized to 100 in August 2007.
40
50
60
70
80
90
100
110
120
130
4/28/2007 8/6/2007 11/14/2007 2/22/2008 6/1/2008 9/9/2008 12/18/2008 3/28/2009 7/6/2009 10/14/2009
BofA ML Global Government ex-U.S.IndexBofA ML U.S. Municipal SecuritiesIndexBofA ML U.S. Corporate Bond Index
BofA ML U.S. ABS & CMBS Index
BofA ML U.S. High Yield Index
BofA ML Euro High Yield Index
BofA ML Corporate Bond ex-U.S. Index
TABLE I The Composition of Corporate Reserves
This table shows the fair value of corporate reserve assets by asset classes. Panel A reports equally-weighted firm-level averages over the entire sample period 2009-2012. Panel B reports aggregate sample-wide values as of 2012, the most recent year in our sample. Safe reserves are defined as assets that fall into the following asset classes: cash, cash equivalents, and U.S. Treasuries. Risky reserves are defined as assets that fall into the following asset classes: other government debt, corporate debt, asset-backed and mortgage-backed securities (ABS & MBS), other debt, equity, and other securities. Each reserve asset is manually assigned into a unique asset class based on hand-collected data from footnotes of annual reports. To compare these values with traditional Compustat data, we also report the size of each asset class relative to the following items (Compustat data items are in parentheses): (1) total book assets (AT), (2) the market value of equity (PRCC_F*CSHO), (3) cash and short term investments from the balance sheet (CHE). Finally, we also provide the size of each asset class relative to the total reserve assets reported in the footnote and balance sheet. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999).
Panel A: Firm-level analysis
Security Amount ($M) Fraction of book assets
Fraction of MV of equity
Fraction of CHE
Fraction of total reserves
Cash and cash equivalents 1,940.14 12.16% 13.26% 89.35% 81.52%
Deposits 153.69 1.22% 0.97% 5.11% 4.71%
Commercial paper 82.55 0.44% 0.39% 1.58% 1.39%
Money market funds 192.41 1.72% 1.25% 7.26% 6.45%
Other 1,511.49 8.79% 10.80% 75.41% 68.97%
U.S. Treasuries 157.37 0.57% 0.49% 2.14% 1.90%
Total safe reserves 2,097.50 12.73% 13.75% 91.49% 83.42%
Other government debt 496.56 1.93% 1.35% 7.60% 5.33%
Municipal 54.00 0.45% 0.28% 1.93% 1.21%
Agency 271.57 1.15% 0.79% 4.13% 2.87%
Other 171.00 0.34% 0.34% 1.54% 1.25%
Domestic 373.52 1.68% 1.13% 6.64% 4.52%
Foreign 123.04 0.26% 0.28% 0.96% 0.81%
Corporate debt 317.09 1.41% 1.15% 5.58% 3.91%
Domestic 298.78 1.37% 1.12% 5.29% 3.66%
Foreign 18.32 0.04% 0.04% 0.29% 0.26%
ABS & MBS 93.20 0.42% 0.31% 2.03% 1.40%
Other debt 57.11 0.19% 0.22% 1.38% 0.90%
Equity 70.06 0.25% 0.29% 2.96% 1.43%
Mutual funds 3.91 0.02% 0.02% 0.14% 0.12%
Other 66.15 0.23% 0.27% 2.82% 1.31%
Domestic 68.37 0.24% 0.28% 2.86% 1.36%
Foreign 1.69 0.01% 0.01% 0.10% 0.07%
Other securities 139.23 0.58% 0.51% 5.90% 3.62%
Total risky reserves 1,173.25 4.79% 3.79% 25.45% 16.58%
Total reserves 3,270.76 17.52% 17.22% 116.93% 100.00%
Panel B: Aggregate Analysis
Security Amount ($M) Fraction of book assets
Fraction of MV of equity
Fraction of CHE
Fraction of total reserves
Cash and cash equivalents 888,500 8.35% 8.09% 69.46% 55.75%
Deposits 57,451 0.54% 0.52% 4.49% 3.60%
Commercial paper 32,071 0.30% 0.29% 2.51% 2.01%
Money market funds 78,063 0.73% 0.71% 6.10% 4.90%
Other 720,914 6.78% 6.56% 56.36% 45.23%
U.S. Treasuries 94,337 0.89% 0.86% 7.37% 5.92%
Total safe reserves 982,837 9.24% 8.95% 76.83% 61.67%
Other government debt 243,675 2.29% 2.22% 19.05% 15.29%
Municipal 27,358 0.26% 0.25% 2.14% 1.72%
Agency 139,336 1.31% 1.27% 10.89% 8.74%
Other 76,981 0.72% 0.70% 6.02% 4.83%
Domestic 178,032 1.67% 1.62% 13.92% 11.17%
Foreign 65,643 0.62% 0.60% 5.13% 4.12%
Corporate debt 173,570 1.63% 1.58% 13.57% 10.89%
Domestic 165,260 1.55% 1.50% 12.92% 10.37%
Foreign 8,309 0.08% 0.08% 0.65% 0.52%
ABS & MBS 56,634 0.53% 0.52% 4.43% 3.55%
Other debt 20,297 0.19% 0.18% 1.59% 1.27%
Equity 33,031 0.31% 0.30% 2.58% 2.07%
Mutual funds 1,612 0.02% 0.01% 0.13% 0.10%
Other 31,419 0.30% 0.29% 2.46% 1.97%
Domestic 31,667 0.30% 0.29% 2.48% 1.99%
Foreign 1,363 0.01% 0.01% 0.11% 0.09%
Other securities 83,781 0.79% 0.76% 6.55% 5.26%
Total risky reserves 610,987 5.75% 5.56% 47.76% 38.33%
Total reserves 1,593,824 14.99% 14.51% 124.60% 100.00%
TABLE II Liquidity of Corporate Reserves
This table shows the percent of each asset class classified by the firm as level 1, level 2, or level 3 assets. Panel A reports equally-weighted firm-level averages over the entire sample period 2009-2012. Panel B reports aggregate sample-wide values as of 2012, the most recent year in our sample. Asset levels are measures of liquidity since they are determined by the type of inputs required to assess the fair value of the asset: quoted prices in active markets for identical assets (level 1), significant other observable inputs (level 2), or significant unobservable inputs (level 3). Thus, level 1 assets are the most liquid, whereas level 3 assets are the most illiquid. Safe reserves are defined as assets that fall into the following asset classes: cash, cash equivalents, and U.S. Treasuries. Risky reserves are defined as assets that fall into the following asset classes: other government debt, corporate debt, asset-backed and mortgage-backed securities (ABS & MBS), other debt, equity, and other securities. Each reserve asset is manually assigned into a unique asset class based on hand-collected data from footnotes of annual reports. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999).
Panel A: Firm-level analysis
Security Level 1 Level 2 Level 3
Cash and cash equivalents 93.71% 6.29% 0.00%
Deposits 61.43% 38.57% 0.00%
Commercial paper 16.00% 84.00% 0.00%
Money market funds 88.20% 11.62% 0.18%
Other 98.26% 1.74% 0.00%
U.S. Treasuries 71.58% 27.09% 1.32%
Total safe reserves 94.13% 5.81% 0.06%
Other government debt 25.99% 72.88% 1.13%
Municipal 10.33% 87.67% 2.01%
Agency 31.34% 68.60% 0.07%
Other 17.27% 81.87% 0.86%
Domestic 44.06% 54.99% 0.95%
Foreign 15.49% 83.65% 0.86%
Corporate debt 9.60% 88.05% 2.35%
Domestic 9.24% 88.37% 2.39%
Foreign 10.97% 79.09% 9.94%
ABS & MBS 3.23% 49.86% 46.91%
Other debt 21.59% 68.96% 9.45%
Equity 86.69% 8.83% 4.49%
Mutual funds 86.70% 13.30% 0.00%
Other 85.90% 8.66% 5.44%
Domestic 86.81% 8.92% 4.27%
Foreign 91.67% 0.00% 8.33%
Other securities 62.82% 26.82% 10.36%
Total risky reserves 43.28% 45.13% 11.59%
Total reserves 86.23% 12.83% 0.94%
Panel B: Aggregate analysis
Security Level 1 Level 2 Level 3
Cash and cash equivalents 89.98% 10.02% 0.00%
Deposits 71.75% 28.25% 0.00%
Commercial paper 4.42% 95.58% 0.00%
Money market funds 92.34% 7.66% 0.00%
Other 95.53% 4.47% 0.00%
U.S. Treasuries 42.27% 57.44% 0.29%
Total safe reserves 86.12% 13.85% 0.03%
Other government debt 32.03% 67.90% 0.06%
Municipal 0.26% 99.40% 0.33%
Agency 42.33% 57.67% 0.00%
Other 20.28% 79.62% 0.09%
Domestic 39.56% 60.30% 0.14%
Foreign 12.25% 87.66% 0.09%
Corporate debt 1.06% 95.37% 3.57%
Domestic 0.98% 96.15% 2.87%
Foreign 4.14% 62.92% 32.96%
ABS & MBS 0.13% 87.78% 12.09%
Other debt 23.14% 75.99% 0.88%
Equity 89.79% 6.85% 3.36%
Mutual funds 92.56% 7.44% 0.00%
Other 89.64% 6.82% 3.54%
Domestic 92.07% 7.15% 0.78%
Foreign 37.47% 0.00% 62.53%
Other securities 20.40% 77.39% 2.20%
Total risky reserves 21.23% 75.97% 2.80%
Total reserves 63.21% 35.78% 1.01%
TABLE III Industry Analysis
This table analyzes the fair value of risky reserves by the Fama and French 5-industry classification (Panels A and B) and lists the top 20 firms on risky reserves as of 2012 (Panel C). Panel A reports firm-level averages over the entire sample period 2009-2012. Panel B reports aggregate values as of 2012, the most recent year in our sample. Risky reserves are defined as assets that fall into the following asset classes: non-US government debt, corporate debt, asset-backed and mortgage-backed securities (ABS & MBS), other debt, equity, and other securities. Each reserve asset is manually assigned into a unique asset class based on hand-collected data from footnotes of annual reports. In Panels A and B, we also report the size of each asset class relative to the following items (Compustat data items are in parentheses): (1) total book assets (AT), (2) the market value of equity (PRCC_F*CSHO), (3) cash and short term investments from the balance sheet (CHE). Finally, we also provide the size of each asset class relative to the total reserve assets reported by the firm. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999). Panel A: Firm-level analysis of risky reserves by industry
Fama-French Industry Amount ($M) Percent of book assets
Percent of MV of equity
Percent of CHE
Percent of total reserves
Consumer 492.35 2.21% 3.07% 20.08% 11.12%
Manufacturing 233.42 1.44% 1.46% 15.87% 10.12%
Technology 2,088.67 11.07% 6.85% 36.25% 26.86%
Health 2,418.28 7.35% 4.45% 42.56% 26.41%
Other 1,681.09 1.88% 3.18% 21.03% 12.34%
Panel B: Aggregate analysis of risky reserves by industry
Fama-French Industry Amount ($M) Percent of book assets
Percent of MV of equity
Percent of CHE
Percent of total reserves
Consumer 65,537.26 3.24% 2.75% 30.74% 27.23%
Manufacturing 28,198.26 0.94% 1.03% 13.62% 12.87%
Technology 322,928.30 11.53% 9.22% 66.61% 50.35%
Health 106,364.10 11.12% 8.22% 63.55% 51.83%
Other 87,959.34 4.75% 8.20% 42.54% 30.60%
Panel C: Top 20 firms on risky reserves in 2012
Rank Risky reserves (millions of dollars) Risky reserves (over total assets) Risky reserves (over total reserves)
Name Amount Name Percent Name Percent
1 APPLE INC 92,271 VERISIGN INC 70.58% MICROSOFT CORP 97.14%
2 MICROSOFT CORP 69,480 ANALOG DEVICES 66.19% LEUCADIA NATIONAL CORP 95.83%
3 GENERAL ELECTRIC CO 45,779 MICROSOFT CORP 57.29% ANALOG DEVICES 95.40%
4 BERKSHIRE HATHAWAY 32,291 INTUITIVE SURGICAL INC 56.30% CF INDUSTRIES HOLDINGS INC 95.39%
5 PFIZER INC 29,314 QUALCOMM INC 54.00% VERISIGN INC 93.26%
6 QUALCOMM INC 23,227 XILINX INC 53.97% MEDTRONIC INC 92.38%
7 GOOGLE INC 21,959 APPLE INC 52.41% TEXAS INSTRUMENTS INC 89.50%
8 GENERAL MOTORS CO 18,056 F5 NETWORKS INC 51.03% WHOLE FOODS MARKET INC 87.44%
9 FORD MOTOR CO 16,543 NVIDIA CORP 48.37% SANDISK CORP 86.11%
10 JOHNSON & JOHNSON 15,945 SANDISK CORP 47.57% QUALCOMM INC 85.92%
11 AMGEN INC 15,637 QLOGIC CORP 43.62% BIOGEN IDEC INC 84.87%
12 CISCO SYSTEMS INC 14,676 LINEAR TECHNOLOGY CORP 41.27% NVIDIA CORP 83.19%
13 INTEL CORP 12,396 LEUCADIA NATIONAL CORP 35.91% HELMERICH & PAYNE 83.09%
14 MEDTRONIC INC 10,499 GARMIN LTD 34.06% HEALTH MANAGEMENT ASSOC 82.19%
15 COCA-COLA CO 9,876 MICROCHIP TECHNOLOGY INC 33.95% F5 NETWORKS INC 81.61%
16 MERCK & CO 8,052 AKAMAI TECHNOLOGIES INC 33.94% AKAMAI TECHNOLOGIES INC 80.59%
17 LILLY (ELI) & CO 7,746 WATERS CORP 33.39% QLOGIC CORP 79.03%
18 EMC CORP/MA 6,642 TELLABS INC 33.18% INTUITIVE SURGICAL INC 78.25%
19 ORACLE CORP 4,935 KLA-TENCOR CORP 32.30% APPLIED MICRO CIRCUITS CORP 77.70%
20 SANDISK CORP 4,918 NETAPP INC 32.12% LEXMARK INTL INC -CL A 77.54%
TABLE IV Corporate Reserves (Scaled by Book Assets) and Traditional Measures of Cash or Excess Cash
This table provides evidence on the relation between corporate reserve assets, as a percentage of book assets, and traditional Compustat-based measures of cash or excess cash. In Panel A, we sort the firms in our sample into annual quintiles based on their Compustat-based measure of cash holdings, defined as cash and short-term investment (CHE) divided by book assets (AT). In Panel B, we sort firms into annual quintiles based on their excess cash holdings, estimated as the regression residuals from regressing their Compustat-based measure of cash holdings on the following set of explanatory variables: cash flow volatility, the market-to-book ratio, cash flow over assets, size, and year and industry fixed effects. In Panel C, we sort firms into annual quintiles based on their total reserves. Safe reserves are defined as assets that fall into the following asset classes: cash, cash equivalents, and U.S. Treasuries. Risky reserves are defined as assets that fall into the following asset classes: other government debt, corporate debt, asset-backed and mortgage-backed securities (ABS &MBS), other debt, equity, and other securities. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999). Panel A: Asset classes and traditional cash holdings
Quintiles based on traditional Compustat-based cash holdings
Security Low 2 3 4 High
Cash and cash equivalents 2.14% 6.05% 10.94% 16.60% 24.96%
U.S. treasuries 0.01% 0.06% 0.18% 0.53% 2.04%
Total safe reserves 2.16% 6.11% 11.13% 17.13% 27.00%
Other government debt 0.12% 0.17% 0.47% 2.02% 6.86%
Corporate debt 0.03% 0.10% 0.67% 1.36% 4.88%
ABS & MBS 0.03% 0.06% 0.13% 0.39% 1.48%
Other debt 0.02% 0.03% 0.21% 0.20% 0.50%
Equity 0.21% 0.21% 0.14% 0.24% 0.43%
Other securities 0.18% 0.25% 0.42% 0.70% 1.35%
Risky reserves 0.59% 0.82% 2.04% 4.92% 15.50%
Panel B: Asset classes and traditional excess cash holdings
Quintiles based on traditional Compustat-based excess cash holdings
Security Low 2 3 4 High
Cash and cash equivalents 6.30% 6.85% 9.64% 13.99% 22.03%
U.S. treasuries 0.16% 0.08% 0.31% 0.37% 1.91%
Total safe reserves 6.45% 6.93% 9.95% 14.37% 23.94%
Other government debt 0.77% 0.76% 0.95% 1.94% 5.48%
Corporate debt 0.68% 0.77% 0.89% 1.03% 4.42%
ABS & MBS 0.17% 0.14% 0.13% 0.43% 1.07%
Other debt 0.06% 0.08% 0.10% 0.30% 0.48%
Equity 0.21% 0.21% 0.15% 0.25% 0.36%
Other securities 0.47% 0.39% 0.26% 0.70% 1.16%
Risky reserves 2.37% 2.34% 2.47% 4.64% 12.97%
Panel C: Asset classes and total reserves
Quintiles based on total reserves
Security Low 2 3 4 High
Cash and cash equivalents 5.68% 12.04% 12.98% 16.22% 13.85%
U.S. treasuries 0.03% 0.13% 0.66% 0.72% 1.29%
Total safe reserves 5.71% 12.17% 13.63% 16.94% 15.14%
Other government debt 0.25% 0.66% 1.95% 2.49% 4.29%
Corporate debt 0.23% 0.46% 1.43% 1.88% 3.05%
ABS & MBS 0.12% 0.29% 0.33% 0.69% 0.68%
Other debt 0.18% 0.07% 0.10% 0.22% 0.40%
Equity 0.09% 0.12% 0.14% 0.32% 0.57%
Other securities 0.16% 0.39% 0.79% 0.79% 0.77%
Risky reserves 1.03% 1.99% 4.74% 6.39% 9.75%
TABLE V Liquidity and Traditional Measures of Cash or Excess Cash
This table provides evidence on the relation between the liquidity of corporate reserve assets, as measured by asset levels, and traditional Compustat-based measures of cash or excess cash. In Panel A, we sort the firms in our sample into annual quintiles based on their Compustat-based measure of cash holdings, defined as cash and short-term investment (CHE) divided by book assets (AT). In Panel B, we sort firms into annual quintiles based on their excess cash holdings, estimated as the regression residuals from regressing their Compustat-based measure of cash holdings on the following set of explanatory variables: cash flow volatility, the market-to-book ratio, cash flow over assets, size, and year and industry fixed effects. In Panel C, we sort firms into annual quintiles based on their total reserves. Asset levels are measures of liquidity since they are determined by the type of inputs required to assess the fair value of the asset: quoted prices in active markets for identical assets (level 1), significant other observable inputs (level 2), or significant unobservable inputs (level 3). Thus, level 1 assets are the most liquid, whereas level 3 assets are the most illiquid. Liquid (illiquid) reserves are defined as level 1 (levels 2 and 3) assets. Risky reserves are defined as assets that fall into the following asset classes: other government debt, corporate debt, asset-backed and mortgage-backed securities (ABS &MBS), other debt, equity, and other securities. Each reserve asset is manually assigned into a unique asset class based on hand-collected data from footnotes of annual reports. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999). Panel A: Asset levels and traditional cash holdings
Quintiles based on traditional Compustat-based cash holdings
Liquidity Low 2 3 4 High
Liquid (level 1) 95.14% 94.63% 90.96% 85.68% 64.91%
Illiquid (levels 2-3) 4.86% 5.37% 9.04% 14.32% 35.09%
Total 100.00% 100.00% 100.00% 100.00% 100.00%
Risky & Liquid (level 1) 59.26% 58.82% 55.76% 41.75% 18.20%
Risky & Illiquid (levels 2-3) 40.74% 41.18% 44.24% 58.25% 81.80%
Total 100.00% 100.00% 100.00% 100.00% 100.00%
Panel B: Asset levels and traditional excess cash holdings
Quintiles based on traditional Compustat-based excess cash holdings
Liquidity Low 2 3 4 High
Liquid (level 1) 90.73% 90.75% 91.14% 85.25% 70.20%
Illiquid (levels 2-3) 9.27% 9.25% 8.86% 14.75% 29.80%
Total 100.00% 100.00% 100.00% 100.00% 100.00%
Risky & Liquid (level 1) 55.90% 55.42% 41.97% 42.98% 25.26%
Risky & Illiquid (levels 2-3) 44.10% 44.58% 58.03% 57.02% 74.74%
Total 100.00% 100.00% 100.00% 100.00% 100.00%
Panel C: Asset levels and total reserves
Quintiles based on total reserves
Liquidity Low 2 3 4 High
Liquid (level 1) 94.24% 94.34% 88.11% 81.94% 72.66%
Illiquid (levels 2-3) 5.76% 5.66% 11.89% 18.06% 27.34%
Total 100.00% 100.00% 100.00% 100.00% 100.00%
Risky & Liquid (level 1) 58.17% 54.74% 48.76% 35.89% 32.14%
Risky & Illiquid (levels 2-3) 41.83% 45.26% 51.24% 64.11% 67.86%
Total 100.00% 100.00% 100.00% 100.00% 100.00%
TABLE VI Summary Statistics
This table provides summary statistics for the sample, which comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999). Cash flow is measured as earnings (EBITDA) less interest and taxes (TXT+XINT) divided by total assets (AT). Cash flow volatility is the 10-year rolling window volatility of cash flow. Market to book is the market value of assets, defined as total assets (AT) minus book equity (CEQ) plus market value of equity (CSHO*PRCC_F) divided by total assets (AT). Foreign income is pretax foreign income (PIFO) divided by total assets (AT). Size is the natural logarithm of the book value of total assets (AT). Leverage is debt in current liabilities (DLC) + long-term debt (DLTT) divided by total assets (AT). Capital expenditure is capital expenditures (CAPX) divided by total assets (AT). R&D expenditure is research and development expense (XRD), set to zero where missing, divided by total assets (AT). Dividend dummy is an indicator equal to 1 if the firm paid cash dividends (DV) that year, and zero otherwise. Net working capital is current assets (ACT) – current liabilities (LCT) divided by total assets (AT). Acquisition expenditure is acquisitions (AQC) divided by total assets (AT). G-Index is an index of shareholder rights shown by Gompers, Ishii, and Metrick (2003) to impact shareholder value. CEO overconfidence is a binary variable equal to 1 if the CEO at some point during his tenure held an option package until the last year before expiration, provided that the package was at least 40% in the money entering its last year. Stock compensation is the ratio of the value of the insider holdings of common stocks of the top 5 executives to their total compensation. Option compensation is the ratio of the value of the stock options of the top 5 executives to their total compensation. CAPM beta is market beta computed from monthly returns, with the CRSP value-weighted index used as the market proxy.
Variable Mean 25th percentile
Median 75th percentile
Standard deviation
Market-to-book 1.965 1.284 1.644 2.287 1.067
Size 9.189 8.359 9.081 9.935 1.190
Cash flow 0.093 0.061 0.088 0.123 0.061
Net working capital 0.021 -0.039 0.019 0.083 0.111
Capital expenditure 0.044 0.018 0.031 0.054 0.043
Leverage 0.250 0.124 0.223 0.336 0.199
Cash flow volatility 0.037 0.017 0.027 0.043 0.035
Dividend dummy 0.676 0.000 1.000 1.000 0.468
R&D expenditure 0.042 0.000 0.006 0.050 0.074
Acquisition expenditure 0.021 0.000 0.001 0.016 0.049
Foreign income 0.039 0.000 0.022 0.060 0.048
G-index 9.277 8.000 9.000 11.000 2.026
CEO age 56.740 52.000 57.000 61.000 6.389
Stock compensation 0.458 0.294 0.453 0.624 0.259
Option compensation 0.273 0.079 0.260 0.398 0.227
CEO overconfidence 0.396 0.000 0.000 1.000 0.489
CAPM beta 1.192 0.828 1.102 1.500 0.538
TABLE VII The Determinants of the Composition of Corporate Reserves
This table reports OLS regression evidence on the determinants of the composition of corporate reserves. Risky reserves are defined as assets that fall into the following asset classes: other government debt, corporate debt, asset-backed and mortgage-backed securities (ABS &MBS), other debt, equity, and other securities. Safe reserves are defined as assets that fall into the following asset classes: cash, cash equivalents, and U.S. Treasuries. Each reserve asset is manually assigned into a unique asset class based on hand-collected data from footnotes of annual reports. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999). The regressions include year and industry fixed effects, which are not shown. The standard errors (in brackets) are heteroskedasticity consistent and clustered at the firm level. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.
Panel A: Baseline model
Dependent variable Compustat cash / assets
Total reserves / assets
Risky reserves / assets
Safe reserves / assets
Risky reserves / total reserves
Model (1) (2) (3) (4) (5)
Safe reserves -0.143***
[0.029]
Risky reserves -0.172***
[0.032]
Market to book 0.030*** 0.036*** 0.020*** 0.023*** 0.023*** [0.004] [0.004] [0.004] [0.003] [0.006]
Size -0.016*** -0.013*** 0.004 -0.018*** 0.038*** [0.003] [0.003] [0.002] [0.002] [0.005]
Cash flow 0.181** 0.175** 0.137*** 0.067 0.415*** [0.074] [0.085] [0.050] [0.065] [0.103]
Net working capital -0.235*** -0.274*** -0.105*** -0.210*** -0.070 [0.029] [0.031] [0.022] [0.028] [0.050]
Capital expenditure -0.552*** -0.650*** -0.211*** -0.535*** 0.121 [0.057] [0.068] [0.046] [0.052] [0.129]
Leverage -0.105*** -0.149*** -0.079*** -0.093*** -0.068* [0.014] [0.017] [0.012] [0.014] [0.035]
Cash flow volatility 0.551*** 0.553*** 0.134 0.500*** -0.016 [0.091] [0.111] [0.082] [0.069] [0.167]
Dividend dummy -0.012** -0.029*** -0.019*** -0.014** -0.040*** [0.006] [0.007] [0.005] [0.006] [0.012]
R&D expenditures 0.717*** 0.838*** 0.645*** 0.330*** 1.323*** [0.083] [0.100] [0.061] [0.080] [0.093]
Acquisition expenditure -0.393*** -0.393*** -0.207*** -0.247*** -0.150 [0.046] [0.065] [0.041] [0.057] [0.108]
Year fixed effects? Yes Yes Yes Yes Yes
Industry fixed effects? Yes Yes Yes Yes Yes
Adjusted R2 0.510 0.520 0.371 0.345 0.236
N_obs 1,727 1,727 1,727 1,727 1,727
Panel B: Augmented model
Dependent variable Compustat cash / assets
Total reserves / assets
Risky reserves / assets
Safe reserves / assets
Risky reserves / total reserves
Model (1) (2) (3) (4) (5)
Safe reserves -0.164***
[0.030]
Risky reserves -0.188***
[0.030]
Market to book 0.022*** 0.029*** 0.017*** 0.017*** 0.020*** [0.003] [0.004] [0.004] [0.003] [0.007]
Size -0.020*** -0.016*** 0.003 -0.021*** 0.040*** [0.003] [0.003] [0.002] [0.002] [0.005]
Cash flow 0.061 0.056 0.102** -0.034 0.453*** [0.071] [0.080] [0.048] [0.063] [0.103]
Net working capital -0.249*** -0.289*** -0.116*** -0.223*** -0.078 [0.029] [0.031] [0.022] [0.028] [0.049]
Capital expenditure -0.535*** -0.635*** -0.221*** -0.521*** 0.068 [0.056] [0.068] [0.047] [0.052] [0.131]
Leverage -0.087*** -0.131*** -0.075*** -0.079*** -0.072** [0.014] [0.017] [0.012] [0.014] [0.034]
Cash flow volatility 0.471*** 0.475*** 0.114 0.439*** -0.075 [0.094] [0.114] [0.082] [0.073] [0.166]
Dividend dummy -0.018*** -0.034*** -0.021*** -0.019*** -0.036*** [0.006] [0.007] [0.005] [0.005] [0.012]
R&D expenditures 0.653*** 0.776*** 0.632*** 0.285*** 1.351*** [0.078] [0.094] [0.059] [0.077] [0.092]
Acquisition expenditure -0.377*** -0.377*** -0.208*** -0.235*** -0.156 [0.044] [0.064] [0.041] [0.056] [0.106]
Foreign income 0.591*** 0.589*** 0.180*** 0.507*** -0.155 [0.070] [0.080] [0.062] [0.057] [0.114]
G-index -0.004*** -0.005*** 0.001 -0.003** 0.002 [0.001] [0.001] [0.001] [0.001] [0.003]
CEO age 0.001* 0.001 0.001 0.001 -0.001 [<0.001] [0.001] [0.001] [0.001] [0.001]
CEO overconfidence -0.006 -0.009 0.009** -0.007** 0.027** [0.005] [0.006] [0.004] [0.003] [0.012]
Stock compensation -0.002 -0.003 0.036*** -0.025* 0.046** [0.013] [0.016] [0.011] [0.014] [0.029]
Option compensation 0.001 0.009 0.052*** -0.032** 0.064** [0.014] [0.016] [0.012] [0.015] [0.032]
Credit ratings -0.001 -0.001 -0.001 0.001 0.002 [0.001] [0.001] [0.001] [0.001] [0.002]
CAPM beta 0.003 0.003 -0.002 0.005 -0.004 [0.005] [0.006] [0.004] [0.005] [0.012]
Year fixed effects? Yes Yes Yes Yes Yes
Industry fixed effects? Yes Yes Yes Yes Yes
Adjusted R2 0.548 0.547 0.381 0.387 0.252 N_obs 1,334 1,334 1,334 1,334 1,334
TABLE VIII The Determinants of the Liquidity of Corporate Reserves
This table reports OLS regression evidence on the determinants of the liquidity of corporate reserves. Liquid reserves are defined as level-1 assets. Illiquid reserves are defined as level-2 or level-3 assets. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999). The regressions include year and industry fixed effects, which are not shown. The standard errors (in brackets) are heteroskedasticity consistent and clustered at the firm level. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.
Dependent variable Illiquid reserves / assets
Liquid reserves / assets
Illiquid reserves / total reserves
Illiquid reserves / assets
Liquid reserves / assets
Illiquid reserves / total reserves
Model (1) (2) (3) (4) (5) (6)
Liquid reserves -0.153*** -0.212*** [0.038] [0.054]
Illiquid reserves -0.180*** -0.227***
[0.041] [0.055]
Market to book 0.025*** 0.015** 0.040*** 0.022*** 0.018** 0.033*** [0.006] [0.006] [0.010] [0.006] [0.007] [0.011]
Size 0.007 -0.023*** 0.040*** 0.002 -0.022*** 0.033*** [0.005] [0.004] [0.009] [0.005] [0.005] [0.011]
Cash flow 0.149** 0.066 0.423*** 0.016 -0.112 0.355* [0.070] [0.085] [0.151] [0.082] [0.092] [0.185]
Net working capital -0.102*** -0.209*** -0.124 -0.121*** -0.196*** -0.148 [0.033] [0.053] [0.083] [0.038] [0.056] [0.105]
Capital expenditure -0.290*** -0.457*** -0.366** -0.231*** -0.409*** -0.244 [0.079] [0.078] [0.180] [0.079] [0.084] [0.175]
Leverage -0.072*** -0.085*** -0.086* -0.062** -0.066** -0.080 [0.017] [0.027] [0.049] [0.025] [0.033] [0.061]
Cash flow volatility 0.046 0.406*** 0.061 0.039 0.392** 0.096 [0.111] [0.152] [0.233] [0.128] [0.174] [0.240]
Dividend dummy -0.026*** -0.005 -0.057*** -0.032*** 0.001 -0.081*** [0.010] [0.008] [0.020] [0.012] [0.010] [0.026]
R&D expenditures 0.584*** 0.377*** 1.237*** 0.740*** 0.415*** 1.543*** [0.114] [0.085] [0.225] [0.111] [0.120] [0.230]
Acquisition expenditure -0.208*** -0.260*** -0.228** -0.285*** -0.227*** -0.412*** [0.045] [0.058] [0.102] [0.053] [0.082] [0.129]
Foreign income 0.254* 0.539*** -0.127
[0.144] [0.132] [0.258]
G-index -0.001 -0.003 0.001
[0.002] [0.002] [0.005]
CEO age 0.001 0.001** 0.001
[0.001] [<0.001] [0.001]
CEO overconfidence 0.019*** -0.008** 0.036***
[0.006] [0.003] [0.011]
Stock compensation -0.005 0.003 -0.016
[0.014] [0.014] [0.033]
Option compensation 0.038** -0.019 0.047***
[0.017] [0.016] [0.012]
Credit ratings 0.001 0.001 0.004
[0.001] [0.001] [0.003]
CAPM beta 0.004 0.003 0.011
[0.005] [0.005] [0.013] Year fixed effects? Yes Yes Yes Yes Yes Yes Industry fixed effects? Yes Yes Yes Yes Yes Yes Adjusted R2 0.374 0.323 0.289 0.427 0.364 0.334 N_obs 1,727 1,727 1,727 1,334 1,334 1,334
Table IX The Sources of Corporate Reserves
This table presents estimates from panel regressions explaining the sources of corporate reserve assets. The dependent variable is the annual change in reserve assets, defined as follows
1 .
Safe reserves are defined as assets that fall into the following asset classes: cash, cash equivalents, and U.S. Treasuries. Risky reserves are defined as assets that fall into the following asset classes: other government debt, corporate debt, asset-backed and mortgage-backed securities (ABS &MBS), other debt, equity, and other securities. Each reserve asset is manually assigned into a unique asset class based on hand-collected data from footnotes of annual reports. Cash flow is net income plus amortization and depreciation. Debt and equity issue is the cash proceeds from debt sales and share issuance. Other is all other cash sources, which include the sales of property and the sale of investments. These measures are scaled by lagged total assets. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999) All regressions include year and industry fixed effects, which are not shown. The standard errors (in brackets) are heteroskedasticity consistent and clustered at the firm level. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.
Dependent variable ΔCompustat cash
ΔTotal reserves
ΔSafe reserves
ΔRisky reserves
Model (1) (2) (3) (4)
Cash flow 0.306*** 0.211** 0.072 0.334*** [0.047] [0.098] [0.060] [0.049]
Debt and equity issues 0.045** 0.042** 0.058*** -0.016 [0.018] [0.021] [0.017] [0.012]
Other 0.057 0.177*** 0.073** 0.101** [0.043] [0.039] [0.034] [0.041]
Year fixed effects? Yes Yes Yes Yes
Industry fixed effects? Yes Yes Yes Yes
Adjusted R2 0.108 0.166 0.088 0.123
N_Obs 1,165 1,165 1,165 1,165
Table X The Value of Corporate Reserves
This table presents estimates from panel regressions explaining annual excess stock returns (columns 1-2), raw returns (columns 3-4), and market-to-book ratios (columns 5-6). In columns 1-2, we estimate the following regression model:
, , ΔTotalreserves , ΔRiskyreserves , ΔE , ΔNA , ΔRD , ΔI , ΔD , Totalreserves , L ,
NF , Totalreserves , ΔTotalreserves , L , ΔTotalreserves , , where , , is a firm’s excess return, defined as the stock’s return during the fiscal year less the return on the matching Fama and French (1993) size and book-to-market portfolio, ΔX indicates a change in X from year t – 1 to t, Ei,t is earnings before extraordinary items, NAi,t is net assets, RDi,t is R&D Expenses, set to zero if missing, Ii,t is interest expenses, Di,t is common dividends, Li,t is leverage, defined as the book debt divided by book debt plus the market value of equity, and NFi,t is net financing (including net equity issues and net debt issues). All variables except leverage and excess return have been deflated by lagged market value of equity of the firm. In columns 3-4, the dependent variable is raw return, and the regressions include annual dummies for the Fama and French (1993) size and book-to-market portfolios. In columns 5-6, we estimate the following regression model:
, Totalreserves , Riskyreserves , E , dE , dE , RD , dRD , dRD , D , dD , dD ,
I , dI , dI , dNA , dNA , dMV , , Where MVi,t is the market value of equity plus the book value of liabilities and dXt indicates a change in X from time t – 2 to t. All variables are divided by net assets. Risky reserves are defined as assets that fall into the following asset classes: other government debt, corporate debt, asset-backed and mortgage-backed securities (ABS &MBS), other debt, equity, and other securities. Total reserves are defined as the sum of risky and safe reserves, where safe reserves include cash, cash equivalents, and U.S. Treasuries. Each reserve asset is manually assigned into a unique asset class based on hand-collected data from footnotes of annual reports. The sample comprises all firms in the S&P 500 index from 2009-2012, excluding financial firms (SIC 6000-6999) and regulated utilities (SIC 4900-4999). The standard errors (in brackets) are heteroskedasticity consistent and clustered at the firm level. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%. Dependent variable Abnormal return Raw return Market-to-book ratio Model (1) (2) (3) (4) (5) (6)
ΔTotal reserves 1.080** 1.072** 1.114** 1.085** [0.456] [0.466] [0.497] [0.508]
ΔRisky reserves -0.218** -0.138* -0.239** -0.158** [0.103] [0.067] [0.115] [0.073]
Total reserves 2.219*** 1.602***
[0.399] [0.424]
Risky reserves -0.514** -0.475**
[0.241] [0.196] Year fixed effects? Yes Yes No No Yes Yes Firm fixed effect? No Yes No Yes No Yes Size-book-to-market-year fixed effects? No No Yes Yes No No Adjusted R2 0.334 0.550 0.449 0.567 0.769 0.979 N_obs 1,071 1,071 1,071 1,071 661 661
Top Related