Competition Type, Competition Intensity, and Financial Policies

Transcription

1 Competition Type, Competition Intensity, and Financial Policies Hyungjin Cho 1 Universidad Carlos III de Madrid Lee-Seok Hwang Seoul National University Current Version: April 2016 Abstract By highlighting the difference between price and non-price competition, we expand prior studies on the relation between product market competition and financial policies such as capital structure and payout decisions. We find that (1) the use of equity financing relative to debt financing increases with non-price competition intensity, and (2) firms accumulate more cash reserves by reducing the distribution of free cash flows to outside investors as the non-price competition intensifies. However, such relations are weaker for firms in price competition industries. We contribute to the literature by providing evidence that the competition type can be an important explanatory factor on the relations between competition intensity and financial policies. 1 Corresponding author. Departamento de Economia de la Empresa, Universidad Carlos III de Madrid, Getafe, Madrid, phone: , chyungji@emp.uc3m.es * We appreciate constructive comments from Bok Baik, Woo-Jong Lee, Seung-Yeon Lim, Kyung-Ho Park, Young Jun Kim and the seminar participants at Seoul National University. 1

2 Competition Type, Competition Intensity, and Financial Policies 1. Introduction This study investigates whether the relation between product market competition intensity and financial policies (i.e., capital structure decision and payout decision) can be explained by the industry-level competition type. Prior studies document that product market competition intensity is the important determinant of firm s behavior. Intense competition reduces profitability and increases the uncertainty of future performance (Gaspar and Massa, 2007; Irvine and Pontiff, 2009). As a response to these negative effects of intense competition, firms reduce their leverage (Ovtchinnikov, 2010; Xu, 2012; Morellec et al., 2014) and accumulate more cash by reducing the payout to equity investors (Hoberg et al., 2014; Morellec et al., 2014). However, those prior studies are silent on whether the type of product market competition (i.e., price or non-price competition) is another determinant of financial policies. Economic studies show that a firm s behavior, as a response to product market competition, can be different between price and non-price competition (Stigler, 1968; Sutton, 1991). Thus, we complement this line of literature by comparing the relation between competition intensity and financial policies under price competition and non-price competition. While the firm under price competition competes by setting lower prices than its competitors, non-price competition requires the firm to obtain market share by building up brand value, improving the quality of products and services, and providing reliable guarantee services. Therefore, firms in non-price competition industries have a higher reliance on intangible investments such as advertisement or research and development (R&D) 2

3 expenditures than those in price competition industries (Sutton, 1991; Karuna, 2007; Chen et al., 2015). This different choice of investments can yield different capital structure decisions and payout decisions. Several studies document that advertising expenditures do not significantly increase future sales, implying that the value relevance of advertising expenses is short-lived (Ali Shah and Akbar, 2008). In contrast, most fixed assets are depreciated over more than 5 years, indicating a longer duration of the increase in firm value from capital expenditures. Furthermore, intangible investments are associated with a higher uncertainty of future performance than capital expenditures (Kothari et al., 2002). Given that intense product market competition forces firms to engage in investments to maintain their competency (Nielsen, 2002; Moretto, 2008), the different investment choice of price and nonprice competition industries can result in different relation between competition intensity and future performance in two types of product market competition. In turn, the differential relation between competition intensity and future performance will have different impacts on capital structure and payout decisions. To test the differential effect of price vs. non-price product market competition intensity on financial policies, we use US firms from 1990 to We first use the Herfindahl-Hirschman Index (HHI) of total assets and the HHI of sales, which are based on Compustat database. Although prior studies widely use these measures, HHIs based on Compustat database do not capture the firm s true competitive environments (e.g., Ali et al., 2014; Karuna, 2007). To address this concern, we use two additional measures of competition intensity from recent studies. We use the HHI based on the Census of Manufactures publications provided by the U.S. Census Bureau, which covers all public and private companies. Our last measure of competition intensity is the product market fluidity from Hoberg et al. (2014). In contrast to HHI measures, the fluidity measure captures the firmlevel product market competition intensity as well as the threat from potential entrants. We 3

4 classify the industries with advertising expenses-to-sales ratios higher (lower) than the industry-level median as non-price (price) competition industries because prior studies report that non-price competition industries have a higher reliance on advertising activities than price competition industries (Stigler, 1968; Sutton, 1991). 1 We start an empirical analysis by showing that the extent that competition intensity reduces future profitability is greater for non-price competition industries than for price competition industries. We then link this finding to capital structure decisions. We find that the negative relation between competition intensity and the use of debt financing relative to equity financing is more pronounced for non-price competition industries than for price competition industries. We also find that the positive relation between competition intensity and a conservative payout policy (i.e., less payout to outside investors) is stronger for nonprice competition industries than for price competition industries. Specifically, firms accumulate more cash from free cash flows and distribute a smaller amount of these free cash flows to debt and equity investors when non-price competition is more intense, whereas such relations are not significant in price competition industries. Further tests show that these findings are robust when we control for firm-specific factors such as growth opportunities and the firm life cycle and when we use the deregulation as an exogenous shock to product market competition. To address the case that different firms within the same industry face different types of product market competition, we construct a firm-level indicator of non-price competition. A firm is assumed to face non-price competition when it has a high advertising expenses-tosales ratio, a high R&D expenses-to-sales ratio, a high market-to-book ratio, low capital expenditures, and a short history. The empirical results show that the negative relations that 1 Overall results remain qualitatively similar when we use the ratio of the sum of advertising expenses and R&D expenses on sales in order to classify the sample into price and non-price competition. 4

5 competition intensity has with the use of debt financing relative to equity financing and with the payout to outside investors are more pronounced for firms with a higher likelihood of facing non-price competition, supporting our previous findings. 2 We also examine the consequences of conservative payout policies under intense product market competition. While an increase in cash holding generally leads to more investments in the future, price competition intensity does not significantly influence the positive relation between the change in cash and future investments. In contrast, the change in cash holding is negatively related with future investments when non-price competition intensity is high. This implies that firms under intense non-price competition accumulate more cash holdings and do not use them to execute investments because they are concerned of negative future outlooks from non-price competition. This study contributes to the literature by shedding light on how the competition type influences the relation between competition intensity and financial policies such as financing choices and payout decisions. Prior studies are silent on the possibility that the relation between competition intensity and firm behavior can be an industry-specific phenomenon. However, given that the extent that product market competition intensity influences future performance can be different depending on the type of product market competition, prior findings should be generalized with caution. Particularly, more negative effect of non-price competition on future profitability than price competition suggests that the impact of product market competition on financial policies can be larger for non-price competition than for price competition. This paper is partly related with Gatchev et al. (2009) who report that advertising expenditures and R&D expenditures are mainly funded by equity financing rather than by 2 Karuna (2007) use the price-cost margin, the amount of industry sales, and industry-level fixed assets to define the price competition intensity. However, his paper focuses on the construction of price competition intensity measure rather than on the comparison between price and non-price competition. 5

6 debt financing. However, they are silent on the possibility that competition affects financing decisions, this study investigates the relation between competition intensity and financing choices conditional on competition types. Furthermore, this paper shows that firms under more intense non-price competition rely more on equity financing than debt financing, even when they fund capital expenditures, whereas such a relation does not hold under price competition (see Appendix B). The remainder of this paper proceeds as follows. Section 2 reviews prior literature and Section 3 explains the research design and sample. Section 4 tests the differential effects of competition intensity on future performance between price and non-price competition industries. The tests on the relation between competition and financing choices are presented in Section 5, and Section 6 show the results on the relation between competition and payout decisions. Section 7 provides additional tests and Section 8 concludes the paper. 2. Literature Review 2.1. Product Market Competition, Financing Choices, and Payout Policies Prior studies investigate the relation between product market competition and firm s behavior. One widely accepted finding is that competition deteriorates profitability and increases the uncertainty of future performance (Gaspar and Massa, 2007; Irvine and Pontiff, 2009; Ovtchinnikov, 2010). The negative impacts of intense competition on future performance motivate the firm to reduce leverage in order to decrease default risks. In addition, a decrease in future free cash flows alleviates the concern that managers will appropriate the firm s resource for private benefits, further reducing the incentives to maintain debt in the capital structure (Jensen, 1986; Ovtchinnikov, 2010; Xu, 2012). Then, given that firms actively adjust their leverage to the implicit or explicit target leverage (Byoun, 2008; Faulkender et al., 2012), firms will move their leverage to the lower levels as 6

7 product market competition intensifies. Furthermore, the value of cash holdings is greater when competition is more intense because the lower profitability and high uncertainty of future performance increase the likelihood of failing to meet the cash payments to lenders and suppliers (Alimov, 2014). Consistent with this argument, Morellec et al. (2014) find that firms increase their cash reserves as product market competition intensifies. Hoberg et al. (2014) also find that intense product market competition decreases the payout to equity holders. Fresard (2010) extends this literature by showing that a large cash holding leads to the future gain of market share, particularly when competitors face tighter financial constraints Price vs. Non-Price Competition In price competition industries, firms provide homogenous products and services. Thus, they should cut the price of their products and services to attract customers from competitors. In contrast, non-price competition refers to when companies distinguish their products and services from competitors by offering products and service of high quality, establishing higher brand quality, and providing better guarantee services. Sutton (1991) highlights a higher reliance on advertising activities in non-price competition industries. He reports that advertising activities establish the barrier to entry which enables incumbent companies to protect their market shares from new entrants (see also Stigler, 1968). Therefore, price competition industry is characterized by more reliance on tangible investments, whereas non-price competition industry is characterized by more intangible investments. We link the different investment choice between price and non-price competition into the literature on the association between competition intensity and financial policies. Capital expenditures build manufacturing and operating facilities which are utilized for a relatively longer period. Consistent with this rationale, many depreciable assets have a useful life that 7

8 exceeds 5 years (Internal Revenue Service, 2015). In contrast, the effects of advertising expenses on future performance are short-lived. By reviewing studies on advertising activities, Clarke (1976) concludes that the duration of the increase in future sales from advertising expenditures is between 3 and 15 months. Other studies find the effect of advertising activities on sales to be insignificant (Ali Shah and Akbar, 2008). Furthermore, Kothari et al. (2002) find that advertising expenses and R&D expenses are associated with more uncertainty of future performance than capital expenditures are. Given that intense product market competition forces the firms to maintain their competency through investments (Nielsen, 2002; Moretto, 2008), the different investment choices will result in different relations between product market competition and financial policies between price and non-price competition. This is because product market competition influences capital structure and payout decisions through its effect on future performance (Ovtchinnikov, 2010; Xu, 2012). Particularly, if the negative impacts of competition intensity on future performance are more pronounced for non-price competition industries than for price competition industries, the negative relations that competition intensity has with the use of debt financing relative to equity financing and with the payouts to outside investors will be stronger for non-price competition industries than for price competition industries. 3. Research Design and Sample 3.1 Measurement of Competition Intensity We use various proxies to measure the intensity of product market competition. The concentration ratio, or Herfindahl-Hirschman Index (HHI), is the most frequently used measure of product market competition intensity. Thus, we use the Compustat Fundamentals Annual database to construct the HHI using the total assets of firms within the 2-digit SIC 8

9 industry (HHI_Asset). We also calculate HHI using the sales of firms in the same database (HHI_Sale). Lower values of the HHIs imply more intense product market competition. However, Karuna (2007) reports the limitations of concentration ratios as an empirical proxy of competition intensity. First, the concentration ratio can yield conflicting findings when more firms in the industry are considered. Given that private firms are not included in the Compustat database, the concentration ratio based on Compustat can overestimate competition intensity (see also Ali et al., 2009). Second, the concentration ratio does not consider the possibility that prospective entrants can be influenced by the extent of industry concentration. To address these limitations, we use alternative measures of competition intensity. As the third measure of competition intensity, we use HHI based on the Census of Manufactures publications provided by the U.S. Census Bureau, which covers all public and private companies (HHI_Census). Since the Census of Manufactures is published in every 5 years, we assume that the HHI values of 1997, 2002, and 2007 are valid for 5 years period centered on 1997, 2002, and 2007 (Ali et al., 2009). For example, we use HHI in the 1997 Census of Manufactures for observations from 1995 to We match the HHI from Census of Manufactures to Compustat data using 3-digit NAICS (North American Industry Classification System) rather than 6-digit NAICS. While using 6-digit NAICS can maximize the cross-sectional variation of product market competition in our sample, it can increase the measurement error in the industry-level advertising expense-to-sales ratio, which is critical in the price and non-price competition partition, due to small number of firms within each industry. In contrast, the use of 3-digit NAICS can reduce the effect of outliers in the industry-level advertising expense-to-sales ratio and increase the accuracy in the price and non-price competition partition. To ease the interpretation, we multiply HHI_Asset, HHI_Sale and HHI_Census with (-1) to make their higher value correspond to more intense 9

10 competition. 3 As the last measure of product market competition, we use the product market fluidity measure (Fluid) from Hoberg et al. (2014). Hoberg et al. (2014) use the business descriptions in 10-K filings to develop the product market fluidity measure, which captures the dynamic structure between the firm s product and those of rival firms. They document that the fluidity measure also addresses the threats from potential entrants. Another advantage of using Fluid is that it captures the firm-level competition intensity, whereas HHI_Asset, HHI_Sale and HHI_Census are estimated at the industry-level. 4 A higher value of Fluid implies tougher product market competition. 5 Several studies use the change in import tariff as a measure of product market competition (Fresard, 2010; Xu, 2012). The advantage of using it is that the change in import tariff is presumably exogenous to the firms behaviors. However, we opt not to use it because the import tariff change can be more closely related with price competition relative to nonprice competition because it shifts the price of imported goods rather than motivates the investments on intangible assets. Consistent with this expectation, untabulated tests show that the change in tariff has significant relations with financial decisions of firms under price competition industries, whereas it does not have significant impact on firms under non-price competition industries. 3.2 Measurement of Competition Type Price and non-price competition industries are classified based on advertising activities (Stigler, 1968; Sutton, 1991). Specifically, the industries with an advertising 3 We divide HHI_Census by 100 to ease the interpretation of coefficients in the regression results. 4 We note that the partition of price and non-price competition based on SIC or NAICS classification in the test using fluidity as the measure of competition intensity can suffer from measurement error because the fluidity measure captures the competition beyond the static industry classification such as SIC or NAICS. 5 Product market fluidity data are available on Gerard Hoberg s website ( 10

11 expense to-sales ratio that are higher (lower) than the industry-level median are classified as non-price (price) competition industries. Since this classification criteria of competition type is at the industry-level, whether the firm is subject to price or non-price competition is exogenous to the firm s investment and financing decisions, reducing the endogeneity concern. We use 2-digit SIC classification when our measure of competition intensity is HHI_Asset or HHI_Sale, and use 3-digit NAICS classification when the competition intensity measure is HHI_Census or Fluid. Note that Fluid is the firm-level measure of product market competition. Thus, using 3-digit NAICS classification for price and non-price competition would bias our results when we use Fluid as the measure of industry competition intensity. We believe that this concern is not critical because we find consistent empirical results when we use various industry definitions such as 3-digit NAICS classification, 2-digit SIC classification, and 4-digit SIC classification as well as firm-level classification of price vs. non-price competition. 3.3 The System of Equations Approach to Capture Financing Choices and Payout Decisions A typical research design implemented to investigate the relation between competition intensity and financing or payout decisions is the unconditional regression of external financing on competition intensity measure. However, this research design can yield biased empirical results due to the omitted variable problem. The cash shortfalls from investment and operating activities are the main driver of external financing (Myers and Majluf, 1974; Kim and Weisbach, 2008). This implies that the omission of cash shortfalls in the regression model results in the omitted variable problem. We thus examine the relation between competition intensity and financing or payout choices conditional on internal cash flows and cash holdings. Furthermore, we use the system of equations to address the interdependence of 11

12 financing activities by imposing the restriction that cash inflows (i.e., changes in cash holdings, debt and equity financing) are equal to cash outflows (financing needs). Based on Gatchev et al. (2009), the construction of a system of equations starts from the restriction that cash inflows are equal to cash outflows as follows: - ΔCash + ΔDebt + ΔEquity = Def (1) where ΔCash is the change in cash holdings, ΔDebt is net debt issue, and ΔEquity is net equity issue. Def is the financing deficit, which captures the firm s financing needs. It is calculated as the sum of capital expenditure, the increase in working capital, acquisitions, and dividend payments, minus cash flows from operations and sales of property, plant, and equipment, all scaled by lagged total assets (Frank and Goyal, 2003). The detailed definitions of variables are in Appendix A. Using this restriction, we construct the system of equations with the interaction term of competition intensity and financing deficit as the independent variable. When Def is positive, the system of equations tests the effect of competition intensity on the interdependence of financing activities to fund financing needs. When Def is negative, the model examines the effect of competition intensity on the distribution of free cash flows to cash reserves and external investors. yi,t+1 = B1 Compi,t + B2 Defi,t+1 + B3 Compi,t*Defi,t+1 + C zi,t + et+1 (2) where y is a 3 x 1 vector of financing choices (i.e., ΔCash, ΔDebt, and ΔEquity). B and C are the 3 x 1, and 3 x k vectors of coefficients on the independent variables, respectively. z is a k x 1 vector of determinants of financing choices, and Comp is the measure of competition intensity. To maintain the accounting identity in Equation (1), we impose the following crossequation restrictions on the coefficients: i'b1 = 01 x 1, i'b2 = 11 x 1, i'b3 = 01 x 1, i'c = 01 x k, and 12

13 i'e = 01 x 1. This implies that competition intensity is related with the association between financing deficit and financing choices, but does not change the restriction in Equation (1). We use one-year-lagged values of Comp to reduce the bias from simultaneity problems because capital structure choices can influence the firm s survival rate and competition intensity. The system of equations is estimated using the maximum likelihood method with standard errors clustered at the firm-level (Gould et al., 2006). 6 We further mitigate correlated omitted variable problem by controlling for firm characteristics that can influence financing choices as well as competition intensity. We control for firm size (Size) as larger firms have more stable cash flows and thus are more capable of attracting debt financing. We include the book-to-market of equity (BM), an inverse measure of growth opportunities, because growth options decrease the underinvestment costs and free cash flows problem, reducing the benefits of using debt financing over equity financing (Barclay et al., 2006). Leverage (Lev) is positively related with equity financing and negatively with debt financing because high leverage increases financial distress. CFVola is the volatility of operating cash flows over at least three of the last five years. Higher volatility is associated with a lower level of investment, reducing the demand for external financing (Minton and Schrand, 1999). We also control for the volatility of past performances using the percentage of years that the firm reports losses in net income in at least three of the last five years (Loss%), because high default risk reduces the optimal level of leverage. Tangibility (Tangible) and depreciation and amortization costs (Dep) are controlled for because fixed assets can be used as collateral for debt financing (Frank and Goyal, 2009). We also include R&D expenses (R&D) and R&D_D, which is an indicator variable that equals one for firms reporting R&D expenses, and zero otherwise. RetVola is the standard deviation of daily stock returns over the fiscal year. Ret is annual stock returns, 6 We find that our findings remain largely unchanged when we use the single regression models. 13

14 included to control for market timing activities of equity financing (Baker and Wurgler, 2002). We estimate the system of equations after partitioning the sample by the sign of financing deficit. Prior studies conventionally assume that financing activities have a linear relationship with financing deficit regardless of the sign of the financing deficit (e.g., Shyam- Sunder and Myers, 1999). However, firms with financing surplus (i.e., negative financing deficit) do not need to obtain the proceeds from external financing. Instead, they have the incentives to distribute the cash to outside investors. Furthermore, Jensen (1986) suggests that firms with sufficient internal cash flows (positive free cash flows) are subject to agency costs, as managers invest free cash flows in low-return projects. To address a potential asymmetric relation between financing deficit and financing activities, we partition the sample into groups of firms with positive financing deficit and groups of firms with financing surplus Sample Description The sample includes US firms with data available from the intersection of Compustat and CRSP over the period of 1990 to Following prior studies, financial firms (SIC codes ) and utilities ( ) are excluded from the sample. We winsorize all continuous variables at the top and bottom 1% to eliminate the effect of outliers. Panel A of Table 1 shows the annual distribution of firm-year observations with data on financing activities and competition intensity. Note that HHI_Asset, HHI_Sales, and HHI_Census are multiplied with (-1) to make their higher values correspond to more intense product market competition. Competition intensity measures show increasing trends over the sample period, indicating that the product market competition has become more intensive over time (e.g., Irvine and Pontiff, 2009). Panel B of Table 1 displays five 2-digit SIC industries with the highest and lowest ratios of advertising expenses to sales. Industries with the lowest advertising expenses-to- 14

15 sales ratios include coal mining (2-digit SIC: 12), nonmetallic minerals except fuels (14), trucking and warehousing (42), special trade contractors (17), and heavy construction except building (16). Industries with the highest advertising expenses-to-sales ratios include educational services (82), miscellaneous retail (59), metal mining (10), personal services (72), and transportation services (47). The last two columns in Panel B of Table 1 show the industry-level mean (ROA) and the industry-level standard deviation of return-on-assets (Std(ROA)). They show that non-price competition industries have a lower profitability and a larger standard deviation of profitability than price competition industries. [Insert Table 1 around here] 4. Competition Intensity, Competition Type, and Future Performance This section examines the effect of competition intensity on future performance. The test results are presented in Table 2. For each measure of competition intensity, we partition the sample into price and non-price competition industries. While the coefficients on Comp are insignificant in the most columns for the subsample of price competition industries, they are significantly negative in non-price competition industries. Untabulated statistics indicate that the differences between the coefficients on Comp are significant at the 5% level except when the measure of competition intensity is HHI_Census. These results imply that the negative impact of competition intensity on future performance is more pronounced for nonprice competition industries than for price competition industries. Since lower profitability is the main channels through which more intense competition is associated with lower leverage and more conservative payout policies (Ovtchinnikov, 2010; Xu, 2012), the results in Table 2 imply that the relations between competition intensity and financial policies would be stronger for non-price competition 15

16 industries than for price competition industries. 7 [Insert Table 2 around here] 5. Competition Intensity, Competition Type, and Financing Choices Table 3 shows the ratio of debt financing to total external financing for the tercile ranks of financing deficit and competition intensity. To eliminate the firms that do not have the incentives to obtain cash from external financing, we only use firms with positive financing deficit (negative free cash flows). Note that HHI measures (HHI_Asset, HHI_Sales, and HHI_Census) are multiplied with (-1) to make higher values correspond with more intense product market competition. In most tables, the ratio of debt financing to total external financing decreases for higher values of competition intensity. This is consistent with prior studies in that product market competition intensity is negatively related with leverage (Ovtchinnikov, 2010; Xu, 2012). More importantly, a negative relation between competition intensity and the use of debt financing relative to total external financing is more pronounced for non-price competition industries than for price competition industries. Specifically, for non-price competition industries, the firm s reliance on debt financing significantly decreases with competition intensity, whereas the ratio of debt financing to external financing shows an unclear pattern for price competition industries. [Insert Table 3 around here] 7 When we use the standard deviation of future profitability for forward 5 years as the measure of future uncertainty (Kothari et al. 2002), we find no significant difference between price and non-price competition industries. When we use the standard deviation of monthly stock returns over the fiscal year as the dependent variable, the results deliver conflicting implications depending on the competition intensity measure. We find that the positive relation between competition intensity and return volatility is stronger for price competition industries than non-price competition industries when the competition intensity is measured using HHI_Asset or HHI_Sale. However, when we use HHI_Census or Fluid, the positive relation between competition intensity and return volatility is stronger for non-price competition industries than for price competition industries. 16

17 Table 4 shows the estimation results of the system of equations using the subsamples of price and non-price competition industries. In each panel, Columns (1) to (3) present the results of price competition sample and columns (4) to (6) show the results of non-price competition sample. There is a sharp difference in price and non-price competition industries regarding the relation between competition intensity and financing choices. In Panel A and B, where the measure of competition intensity is HHI_Asset or HHI_Sale, the coefficients on Def*Comp are statistically insignificant for price competition industries (industries with advertising expenses-to-sales ratios lower than the median). In contrast, the coefficients on Def*Comp are significant for non-price competition industries in each panel. Particularly, the coefficients on Def*Comp are significantly negative when the dependent variable is debt financing and significantly positive for equity financing. This indicates that firms in non-price competition industries fund investments using relatively more equity financing than debt financing as the product market competition becomes more intense. The positive coefficients on Def*Comp in cash change regression using non-price competition indicate that firms use less cash holding to fund financing deficit as non-price competition intensifies. The bottom of each panel presents the difference in coefficients on Def*Comp between two samples. The differences in the coefficients on the interaction of financing deficit and competition intensity are statistically significant except in Panel C, suggesting that the relation between competition intensity and financing choices to fund investments is significantly different between price and non-price competition industries. 8 This finding is economically significant. For instance of the result using Fluid, the 8 We follow Clogg et al. (1995) to test whether the differences between the coefficients from the two regression models are statistically significant. The z-statistic is calculated as z = (bg1 bg2) /, where bg1 (bg2) and SE(bG1) (SE(bG2)) refer to the coefficient on the variable of interest and its standard errors in the first (second) regression, respectively. 17

18 average firms in the non-price competition industry use $0.049 of cash holding (= *6.288), issue $0.548 of debt (= *6.288), and issue $0.403 of equity (= *6.288) to fund a dollar of financing deficit. Under non-price competition, firms with the third quartile of Fluid spend $0.008 (= *8.9) of cash holding, issue $0.472 of debt (= *8.9), and issue $0.520 of equity (= *8.9) to fund a dollar of financing deficit, whereas those with the first quartile of Fluid use $0.081 (= *4.32) of cash holding, issue $0.605 of debt (= *4.32), and issue $0.314 of equity (= *4.32). Thus, the change of non-price competition from the first quartile to the third quartile increases the use of equity financing by $0.206 (= ), reduces the use of debt financing by $0.133 (= ), and decreases the use of cash holding by $0.073 (= ) when the firm experience an increase of financing deficit by one dollar. [Insert Table 4 around here] One can raise a concern that the results in Table 4 could be attributable to the difference in investment choices between price and non-price competition industries because intangible investments have lower collateral values than tangible investments. We already, at least partly, mitigate this concern by using tangibility (the ratio of fixed assets on total assets) and the depreciation expense (scaled by lagged total assets) as the control variable. To further mitigate this concern, we replace financing deficit in Table 4 with capital expenditures. This reduces the possibility that the difference in collateral values between tangible and intangible investments drives financing choices because the tangibility of capital expenditure is unlikely to vary over the competition type. The results in Appendix B show that there is a significant difference in financing choices on how to fund capital expenditures between price and nonprice competition industries. For price competition industries, the interaction of competition 18

19 intensity and capital expenditures does not have a significant effect on financing choices. However, capital expenditures under more intense non-price competition are associated with larger equity financing, whereas there is no significant relation with debt financing. This indicates that the difference in financing choice between price and non-price competition industries is attributable to the different natures of product market competition rather than investment choices. 6. Competition Intensity, Competition Type, and Payout Decisions This section investigates the effect of competition type on the relation between competition intensity and payout decisions. Tables 5 to 7 examine cash holdings, debt repayment, and equity repurchase, respectively. Cash holdings, debt repayment, and equity repurchase are scaled by financing surplus (Surp, positive free cash flows), rather than by other common variables (e.g., total asset or sales) to show how cash inflows from operating and investment activities are allocated to cash reserves, debt investors, and equity investors. We also present the estimation results of the system of equations in Table 8. In Tables 5 to 8, we only use firms with financing surplus (i.e., negative financing deficit). Table 5 presents the ratio of the change in cash on financing surplus for the tercile ranks of financing surplus and each competition intensity measure. Except when the competition measure is Fluid, the ratio of the change in cash holdings on financing surplus does not significantly change over the tercile of competition intensity for price competition industries. In contrast, it significantly increases with competition intensity in non-price competition industries for every measure of competition intensity. This indicates that the positive relation between competition intensity and cash accumulation is stronger for nonprice competition industries than for price competition industries. [Insert Table 5 around here] 19

20 Table 6 shows the univariate test of debt repayment using firms with positive financing surplus. Except when the competition intensity measure is HHI_Census, firms in intense product market competition industries repay a smaller amount of debt than firms that face less competition. Furthermore, the negative relation between debt repayment and competition intensity is more pronounced for non-price competition industries than for price competition industries. For instance, when the competition intensity, captured by Fluid, increases from the first tercile to the third tercile, the ratio of debt repayment on financing surplus decreases by (from to 0.092) for price competition industry firms with large financing surplus. In contrast, it reduces by (from to ) for non-price competition industry firms with large financing surplus. [Insert Table 6 around here] In Table 7, we present the univariate test of net equity repurchase using firms with positive financing surplus. Net equity repurchase is calculated as gross equity repurchase minus new equity issuance, all scaled by financing surplus. Price competition industries do not show a clear relation between equity repurchase and competition intensity. The net repurchase-to-financing surplus ratio increases with the intensity of price competition when the competition intensity is measured by HHI_Sale (Panel B) but decreases when the competition intensity is measured by Fluid (Panel D). In contrast, the results using non-price competition industries largely support more conservative payout policy for firms under more intense non-price competition. Except when we test the firms with small or middle levels of financing surplus using HHI_Census, equity repurchases decrease with competition intensity under non-price competition. Collectively, the descriptive analyses in Tables 5 to 7 show that firms under more intense non-price competition accumulate more cash from free cash flows, 20

21 and distribute less amounts of free cash flows to debt and equity investors, whereas such relations are weaker for firms under price competition. [Insert Table 7 around here] We then estimate the system of equations to examine the relation between competition intensity and the distribution of financing surplus to cash reserves and external investors. Table 8 presents the estimation results. Columns (1) to (3) show the results of price competition industries and columns (4) to (6) show the results of non-price competition industries. To ease the interpretation, we use Surp, the financing surplus or the positive free cash flows from operation and investment activities, rather than Def, which is the negative free cash flows. For instance, in each subsample, the coefficients on Surp are positive when the dependent variable is the change in cash holding, and they are negative when the dependent variable is net debt or net equity financing. This indicates that firms accumulate cash (the positive value of ΔCash), repay debt (the negative value of ΔDebt), and repurchase equity (the negative value of ΔEquity). Except when we use Fluid, price competition intensity does not have a significant relation with the distribution of financing surplus to cash reserves and external investors. In contrast, the most coefficients on Surp*Comp are statistically significant for non-price competition industries. Except when the competition intensity is measured by HHI_Census, the positive coefficients on Surp*Comp in the change in cash regressions indicate that firms accumulate more cash reserves from free cash flows as non-price competition intensifies (Column (4)). Moreover, the coefficients on Surp*Comp are significantly positive in debt and equity financing regressions, suggesting that non-price competition industries distribute less amounts of free cash flows to debt and equity investors when product market competition is more intense (Columns (5) and (6)). 21

22 [Insert Table 8 around here] Prior studies document that firm characteristics such as growth opportunities or firm life cycle influence the financing choices between debt and equity financing as well as the payout policies. Growth opportunities are negatively associated with leverage because these opportunities decrease underinvestment and free cash flow problems, thus reducing the benefits of debt financing (Barclay et al., 2006). The greater need for investments to capture growth opportunities also increases the value of cash reserves and motivates the firm to reduce payouts to outside investors (Opler et al., 1999). Also, young firms are expected to have more growth options than old firms. Thus, similar to firms with high growth opportunities, young firms are more likely to have lower leverage and be more conservative in payout policies than old firms. We investigate whether the stronger relations between competition intensity and financial policies in non-price competition industries, compared to price competition industries, are attributable to growth opportunities or firm life cycle rather than the different nature of price and non-price competition. We use the market-to-book ratio of equity (M/B) as the proxy of growth opportunity (Barclay et al., 2006). Following Collins et al. (2014), we capture life cycle by the combined Z_Score = Z_Sale_GR Z_AGE + Z_CAPEX Z_SIZE, where the Z-variable is calculated by subtracting the mean from the observation and dividing it by the standard deviation for each variable. We classify firms in the lowest (highest) tercile of M/B as firms with low (high) growth opportunities and those in the lower (highest) tercile of Z_Score as old (young) firms. M/B and Z_Score are ranked for the subsample of price and non-price competition industries separately. 9 9 Alternatively, we rank M/B or Z_Score before partitioning the sample into price and non-price competition industries to reduce the concern that M/B or Z_Score is larger for the industries with one competition type than 22

23 In Table 9, Panel A and B (C and D) show the test results using firms with financing deficit (financing surplus) with the partitions based on growth opportunity and life cycle, respectively. Panel A and B (C and D) use firms with financing deficit (surplus) only. Since the results using other measures of competition intensity are similar, we present the test results using HHI_Asset as the measure of competition intensity. In Panels A and B, the coefficients on Def*Comp are largely insignificant for price competition industries regardless of the level of growth opportunities or the life cycle. In contrast, the coefficients on Def*Comp are statistically significant in the regressions using non-price competition industries except for the subsamples of firms with low growth opportunities or old firms. In Panels C and D, for price competition industries, the coefficients on Surp*Comp are significant only when the firms have the middle level of growth opportunities. Among firms in non-price competition industries, the coefficients on Surp*Comp are significant for the firms have high growth opportunities and young firms. These results deliver two implications. First, growth opportunities and the firm s life cycle are the important determinants of financial policies such as financing and payout decisions (Barclay et al., 2006; DeAngelo et al., 2006). Second, more importantly, growth opportunities and life cycles do not explain the difference between price and non-price competition industries. They have strong influences on financing and payout decisions only for non-price competition industries. [Insert Table 9 around here] To further mitigate the concern of correlated omitted variables, we control for several firm characteristics additionally. First, several studies report that the use of debt financing increases with the tax benefits of interest payments. To control for the effect of tax other industries with the different competition type. The results remain largely unchanged (untabulated). 23

24 benefits on debt financing, we include effective tax rate as a variable in the system of equations. The effective tax rate is calculated as the total tax expenses scaled by pretax income or the tax paid scaled by pretax income. Regardless of which effective tax rate is measured, our previous findings remain largely unchanged (untabulated). Second, prior studies document that earnings quality is an important determinant of financing choices (e.g., Chang et al., 2009; Chen et al., 2013). Furthermore, there is an ongoing debate on the relation between competition intensity and earnings quality. Several studies argue that competition intensity is positively related with earnings quality because firms in highly concentrated industries tend to avoid the attention of competitors or politicians by deteriorating information environments (Cheng et al., 2013). Other studies report a negative relation between competition intensity and earnings quality based on the argument that intense competition increases proprietary costs related to the disclosure of high-quality information (Ali et al., 2014). By combining these arguments, Guo et al. (2014) document an inverted U-shape relation between competition and earnings quality. To address this concern, we control for accruals quality in Dechow and Dichev (2002) as an additional control variable in the model. Untabulated results show that our previous findings remain robust after controlling for accruals quality. 7. Additional Tests 7.1 The Firm-Level Identification of Competition Type The main empirical analyses of this paper use the industry-level advertising expenses-to-sales ratio to identify price and non-price competition. Using an industry-level identification of price and non-price competition ignores the possibility that firms within same industry face different types of product market competition. For instance, although several airline companies compete for expensive first- and business-class passengers, other 24

25 low-cost carriers provide cheaper flight services. To address this concern, we develop the firm-level indicator of competition type using the Z_Score of the following five firm characteristics. (1) Advertising expenses-to-sales: As explained earlier, a higher advertising expenses-to-sales ratio is a strong indicator of non-price competition. (2) R&D expenses-to-sales: Firms under non-price competition will spend more resources on R&D expenditures than firms under price competition to develop high quality products and services. (3) Market-to-book ratio: Firms under non-price competition will have higher market-to-book ratios than firms under price competition because spending on intangible assets is an expense item in the income statement, whereas tangible investments are capitalized as assets in the balance sheet. (4) Capital expenditure: Firms under non-price competition will rely less on capital expenditures than firms under price competition due to their higher reliance on intangible expenditures. (5) Firm age: Firms use intangible investments to establish barriers against potential entrants (Sutton, 1991). Thus, young firms are more likely to face non-price competition than old firms. Using these five variables, we construct the Z_Score as follows: Z_Score = Z_AD + Z_R&D + Z_MB Z_CAPEX - Z_AGE. We classify firms with a Z_Score higher (lower) than the third (first) quartile as those facing non-price (price) competition. Consistent with our prediction, firms with a higher value of Z_Score have a higher ratio of advertising expenses-to-sales (1.55%) than firms with a lower Z_Score (0.57%). Table 10 shows the estimation results of the system of equations using the 25