![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
Growth in the cash holdings of US nonfinancial corporations (NFCs) has received considerable attention in recent years. These cash holdings constitute a primary component of the growth in firm-level financial asset holdings often highlighted in analyses of the ‘financialization’ of NFCs. In this article, I use a panel of US corporations to empirically analyze two links between corporate cash holdings and the literature on financialization. First, I find a small but positive relationship between cash holdings and shareholder value ideology among large corporations. I capture the growing entrenchment of shareholder ideology using average industry-level stock repurchases, to proxy for industry-level norms encouraging managers to target stock price-based indicators of firm performance. Second, I find a positive relationship between a firm's cash holdings and a measure of the differential it earns between interest income and expense. Given that cash is classified with short-term marketable (and, therefore, interest-bearing) securities on firm balance sheets, this result lends empirical support to the hypothesis that traditionally nonfinancial firms are increasingly engaged in borrowing and lending for profit.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 These statistics are across-firm yearly medians. Unless otherwise noted, the data in this article is from Compustat; details on the data are in Section Three.
2 Cash and short-term investments are identified in Compustat as item #1, and total current receivables by item #2. This grouping of ‘other’ financial assets includes all other current assets (less inventories) (item #68); investments and advances (#31 and #32); and other (non-current) assets (#69). Note that intangibles are not included. Because (a) and (b) plot yearly medians, the financial asset component series are not additive. See Section Three for additional details on the data and sample.
3 ExecuComp includes firms in the S&P 1500 and so covers a smaller sample of larger firms than Compustat (which is used in Figures 1–3), which includes all US nonfinancial corporations. Again, only nonfinancial corporations are included in this figure.
4 Dodd–Frank reduces this advantage for firms with ‘systemically important’ capital divisions, although—among nonfinancial firms—this designation has only been applied to GE. By being designated as systemically important in 2013, GE Capital became subject to regulation as a bank holding company; following the sales of large segments of its capital division, this status was rescinded in 2016 (https://www.treasury.gov/initiatives/fsoc/designations/Documents/GE%20Capital%20Public%20Rescission%20Basis.pdf).
5 GE is, however, exceptional and—prior to 2008—was one of only a few firms with this triple-A rating, suggesting GE had a greater capacity to expand into financial activities than other NFCs. GE was also downgraded post-crisis (by S&P in 2009 and by Moody's in Citation2012).
6 Data limitations constrain empirical definitions of total financial profits, and one's ability to link components of these financial flows to specific asset categories. Of particular note are difficulties in identifying income from capital gains in firm-level data. As such, financial profits are defined in different ways in the existing literature. Orhangazi (Citation2008), for example, defines financial earnings as the sum of interest income and equity in net earnings; Demir (Citation2009) uses non-operating income.
7 A firm's sources of funds include profits (net of operating expenses, like wages and salaries) earned on fixed capital (which depend on the profit rate and the current stock of fixed capital); profits earned on financial assets (which depend on the financial profit rate and the stock of financial assets); and new funds acquired either from new equity issues or new borrowing. A firm's uses of funds can take the form of capital investment, the acquisition of new financial assets, shareholder payouts (dividend payments or repurchases) or interest payments on debt. Together, these variables form the basis of the specification describing a firm’s holdings of liquid financial assets.
8 Industry-level norms may affect managerial behavior through compensation. CEO compensation is frequently linked to a performance target relative to peer-group benchmarks (Bizjak, Lemmon, and Naveen Citation2008; Frydman and Jenter Citation2010), with industry comprising one benchmark.
9 As such, this variable does not measure the relationship between a firm's independent decision to repurchase stock (or to allocate more or fewer funds to shareholder payouts) and that firm’s decision to hold cash. Rather, the variable measures the relationship between norms that pressure managers to target ‘shareholder value’ and the decision to hold cash. Notably, the results are robust in terms of replacing average industry-level repurchases with a measure of total firm-level shareholder payouts (the sum of buybacks and dividends). These are shown in Column 1 of in the Appendix). Note that, in contrast to total shareholder payouts, the lumpiness of stock repurchase plans implies that firm-level repurchases offer insufficient variation to include independently of dividends.
10 Whereas average industry-level repurchases are strictly exogenous to the individual firm, it is important to note that Gap is predetermined with respect to the decision to hold cash. Accordingly, the discussion here presents controlled correlations consistent with the behavioral hypothesis raised in the article; namely, that one determinant of increased cash holdings among US NFCs is increasing involvement in providing financial services. However, the specification here cannot fully disentangle whether the mechanism runs from financialization to cash holdings or from cash holdings to interest income.
11 Measuring both interest income and expense relative to total assets ensures the measures are comparable. Relative to total assets, interest income and interest expense have an intuitive interpretation—average total interest income and expense per unit of a firm's assets. Ideally, both interest income and expense would be normalized by the respective stock of assets or liabilities that they earn or incur interest on—i.e., firm-specific financial profit rates and interest rates. While it is possible to measure a firm's average effective cost of borrowing, because total debt is known, the specific assets that earn interest income are unknown. Specifically, the hypothesis here considers the possibility that firms earn interest on short-term asset holdings (i.e., their ‘cash’); however, firms may also earn interest on an (unidentifiable) subset of ‘other’ financial assets (see also Section 2.1 and footnote 8).
12 The measure of Gap in Equation 1 is underestimated. However, these other relevant uses of funds are controlled for in both Equations 1 and 2 (where, given debt and capital, outstanding equity is a residual). Furthermore, given that total interest expense reflects debt used to cover all uses of borrowed funds, Gap need not be positive in any given year for borrowing and lending activities to be profitable.
13 The choice of a fixed-effects model over a random-effects model is supported by the Hausman test. A joint test for the significance of the time dummies also confirms the choice to include year fixed effects in the model.
14 The results are robust to measuring Gap as interest income less interest expense, relative to net assets (thereby mirroring the definition of the dependent variable).
15 Tables A1–A3 in the Appendix present sensitivity analysis. successively drops explanatory variables from the specification to show that the main results are insensitive to the inclusion of any single control. Columns 2 and 3 of test for sensitivity to transitory cash holdings—i.e., the possibility that a firm's cash holdings in one year reflect preparation for planned expenditures the following year—by introducing next year's change in cash holdings (the first lead of the first difference) as an explanatory variable (Opler et al. Citation1999, p. 29). , which shows results for the full sample and the largest quartile of firms, indicates that the results are robust to controlling for transitory cash holdings. analyzes robustness relative to additional controls—firm-level R&D spending; a measure of market leverage (defined as total debt relative to the market value of the firm); cash flow (relative to total assets); and an indicator variable defining whether the firm pays a dividend. Definitions of these variables are in the footnote of .
16 Standardized coefficients define the standard deviation of a change in the dependent variable associated with a one standard deviation change in the independent variable. Economic magnitudes in standardized terms, therefore, depend both on the estimated coefficient and on the standard deviation of the dependent and independent variable, for each sample. As such, a decline in the estimated, non-standardized coefficient can be consistent with an increase in the economic magnitude of the estimated relationship (when the standard deviation of the dependent variable relative to that of the independent variable increases). Given that the magnitude of non-standardized estimates depends in part on the mean and variance of the independent variable (which can change across subsamples), standardized coefficients are used here to facilitate interpretation of economic magnitudes across samples. Standardized terms ease comparisons of magnitudes across independent variables by presenting each in terms of a one standard deviation change in the independent variable.
17 It is relevant to note that possible collinearity among the explanatory variables would inflate standard errors on the collinear variables (although without consequences for the point estimates themselves).
18 Section 3.1.2 posits a negative short-term relationship, but a positive relationship between capital and financial asset holdings over time. To consider this distinction, Columns 4 and 5 of introduce two additional lags of the capital stock for the full sample and the largest quartile of firms. These results indicate that the negative coefficient in is limited to the first lag.
19 As one would expect from , coefficients on additional lags of average industry-level repurchases for the three smaller quartiles of firms (not reported in ) are insignificant.