Precautionary Savings with Risky Assets: When Cash Is Not Cash

Transcription

1 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. duchin@uw.edu; gilbertt@uw.edu; jarrad@uw.edu; hrdlicka@uw.edu. 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 9 th 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 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. 2

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 reserves 2 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 covary 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. 3

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 firmby-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 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. 4

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 valueweighted 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. 5

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%). 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. 7

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 8

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 $ at the end of the year to distribute to shareholders. After paying dividend taxes on the distribution, shareholders can expect $ 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 taxinduced 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 9

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. 10

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 11

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. 12

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, reachingfor-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). 13

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. 14

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 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. 15

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. 16

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. 17

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 longterm 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. 18

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. 19

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. 20

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 ( ), 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 21

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. 22

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 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 23

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 % of their total assets and % of their total reserves in risky reserves. As before, Panel B demonstrates that the 24

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 25

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. 26

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 27

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 28

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, 29

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 30

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: 31

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 32

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

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 34

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 35

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. 36

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 37

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 38

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. 39

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:,, ΔTotal reserves, ΔRisky reserves, ΔE, ΔNA, ΔRD, ΔI, ΔD, Total reserves, L, NF, Total reserves, ΔTotal reserves, L, ΔTotal reserves,, 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, E i,t is earnings before extraordinary items, NA i,t is net assets, RD i,t is R&D Expenses, set to zero if missing, I i,t is interest expenses, D i,t is common 40

41 dividends, L i,t is leverage, defined as the book debt divided by book debt plus the market value of equity, and NF i,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:, Total reserves, Risky reserves, E, de, de, RD, drd, drd, D, dd, dd, I, di, di, dna, dna, dmv,, where MV i,t is the market value of equity plus the book value of liabilities and dx t 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. 41

42 The inferences using the alternative approach in columns 5-6 are similar. The point estimates suggest that the value of risky reserves is % 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 42

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. 43

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 Baumol, W. J., 1952, The Transactions Demand for Cash: An Inventory Theoretic Approach, Quarterly Journal of Economics, Vol. 66, pp Bebchuk, L., A. Cohen, and A. Ferrell, 2009, What Matters in Corporate Governance? Review of Financial Studies, Vol. 22, pp 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 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 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 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

45 Duchin, R., 2010, Cash Holdings and Corporate Diversification, Journal of Finance, Vol. 65, pp 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 Fama, E., 1980, Agency Problems and the Theory of the Firm, Journal of Political Economy, Vol. 88, pp 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 Gompers, P., J. Ishii, and A. Metrick, 2003, Corporate Governance and Equity Prices, Quarterly Journal of Economics, Vol. 118, pp Gormley, T. and D. Matsa, 2014, Common Errors: How to (and Not to) Control for Unobserved Heterogeneity, Review of Financial Studies, Vol. 27, pp Harford, J., 1999, Corporate Cash Reserves and Acquisitions, Journal of Finance, Vol. 54, pp Harford, J., S. Mansi and W. Maxwell, 2008, Corporate Governance and Firm Cash Holdings, Journal of Financial Economics, Vol. 87, pp Holmstrom, B., 1999, Managerial Incentive Problems: A Dynamic Perspective, The Review of Economic Studies, Vol. 66, pp Jensen, M., 1986, Agency Costs of the Free Cash Flow, Corporate Finance and Takeovers, American Economic Review, Vol. 76, pp Jensen, M. and W. Meckling, 1976, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics, Vol. 3, pp 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 Kim, W. and M. Weisbach, 2008, Motivations for Public Share Offers: An International Perspective, Journal of Financial Economics, Vol. 87, pp 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. 45

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 Malmendier, U. and G. Tate, 2005, CEO Overconfidence and Corporate Investment, Journal of Finance, Vol. 60, pp McLean, D., 2011, Share Issuance and Cash Savings, Journal of Financial Economics, Vol. 99, pp Milbradt, K., 2012, Level 3 Assets: Booking Profits and Concealing Losses, Review of Financial Studies, Vol. 25, pp Miller, M. and D. Orr, 1966, A Model of the Demand for Money by Firms, Quarterly Journal of Economics, Vol. 80, pp Opler, T., L. Pinkowitz, R. Stulz and R. Williamson, 1999, The Determinants and Implications of Cash Holdings, Journal of Financial Economics, Vol. 52, pp Sufi, A., 2009, Bank Lines of Credit in Corporate Finance: An Empirical Analysis, Review of Financial Studies, Vol. 22, pp

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 largecap 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. 47

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, After a year of transition, it became fully effective for fiscal year 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 48

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). 49

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. 50

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 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 Using this discount rate, the present value of the difference between the firm investing and the individual investing is 51

52 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 1 1. 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; 52

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. 53

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 And the future value of the firm delaying repatriation and investing the reserves itself is The difference in future value between the firm investing and the individual investing is 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: 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). 54

55 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. 55

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. 56