Related Party Transactions, Investments and External Financing. Avishek Bhandari University of Wisconsin - Whitewater

Related Party Transactions, Investments and External Financing Avishek Bhandari University of Wisconsin - Whitewater Mark Kohlbeck * Florida Atlantic University Brian Mayhew University of Wisconsin - Madison PRELIMINARY - DO NOT CITE September 28, 2017 * Corresponding Author School of Accounting 777 Glades Road Boca Raton, FL 33431 562-297-1363 mkohlbec@fau.edu

Related Party Transactions, Investments and External Financing Abstract We investigate the association of related party transactions (RPTs) classified as tone and business with a firm s capital allocation efficiency. We expect that tone RPTs (opportunistic) but not business RPTs (efficient contracting) are associated with inefficiencies in capital allocation, that is, less than optimal investment levels and external financing. We examine investment sensitivity and external financing sensitivity to both Tobin s Q and internally generated cash flows. We provide evidence that tone RPTs, but not business RPTs are associated with distorted investment sensitivity to Q, higher investment sensitivity to internally generated cash flows, lower external financing sensitivity to Q, and higher external sensitivity to internally generated cash flows. Further, we provide evidence consistent with the inefficiencies associated with tone RPTs result in lower future firm accounting and share performance. Our study contributes to the ongoing debate on the costs and benefits of RPTs and provides evidence that all RPTs are not the same. Key Words: Related party transactions, Capital allocation, Investment efficiency, External financing Data Availability: Data is obtained from the publicly-available sources listed in the paper.

Related Party Transactions, Investments and External Financing INTRODUCTION We investigate the association between related party transactions (RPTs) and a firm s capital allocation efficiency. Recent research suggests that RPT type can signal whether insiders appear to use their power to engage in higher levels of opportunism or the transaction is the result of efficient contracting (Kohlbeck and Mayhew 2017). We expect that opportunist RPTs, due to negative effects such as rent extraction, consumption of private benefits, and entrenchment, are associated with inefficiencies in capital allocation, that is, less than optimal capital allocation levels. We examine capital allocation in terms of investments and external financing. First, a high marginal Q firm (that is, a firm with a relatively higher Tobin s Q) is expected to have higher investments. However, there is the potential for opportunistic RPTs to redirect amounts to related parties rather than investments. In this situation, investments will be more dependent on internally generated cash. We posit two reasons to justify the positive effect on the sensitivity of investment to cash flow. First, concerns over agency issues associated with RPTs could result in financiers restricting access to external financing. This restriction in external capital results in investments being more dependent on internally generated cash. Second, entrenched managers of opportunistic RPT firms could try to shield themselves from disciplinary forces of external capital markets by relying on internally generated cash for investments. Second, a high marginal Q firms should also raise more capital as they invest more. However, distortions in investment sensitivities to Q associated with opportunistic RPTs discussed above can manifest themselves in similar effects on the external financing sensitivities. Specifically, opportunistic RPTs are expected to be associated with a detrimental effect on the

external financing-q relationship. In addition, opportunistic RPT activities are many times viewed as strategies that maximize managerial wealth and satisfaction at the expense of investors wealth. Consequently, financing would be more dependent on internally generated cash for these firms. We regress both investments and external financing on Tobin s Q and internally generated cash flows where the estimated coefficients provide measures of investment and external financing sensitivity. We then follow Kohlbeck and Mayhew (2017) and classify RPTs as tone and business where tone RPTs are considered opportunistic and business RPTs a form of efficient contracting. We expect that investments and external financing of tone RPT firms are both less sensitive to Tobin s Q but more sensitive to internally generated cash flow. We also expect these inefficiencies in capital allocation are manifested in lower future firm performance. We find evidence consistent with firms that disclose tone RPTs, but not business RPTs, constraining the sensitivity of current growth opportunities to future investment. That is, the evidence is consistent with tone RPT firms distorting firm-level capital allocation efficiency. We further show that investment is more dependent on internally generated cash for firms with tone RPT reporting firms. We also document that tone RPTs are inversely associated with the external financing sensitivity to Q. In addition, we provide some evidence that firms reporting tone RPTs rely more on internally generated cash for financing, i.e., cash flow is more positively associated with external financing for firms with tone RPTs. Finally, we provide results that are consistent with the prediction that RPT type has important implications for future firm performance. Tone RPTs are associated with lower future accounting and stock performance but business RPTs are not. These results are robust to alternative model specifications as well as variable measurements. 2

Our study makes several contributions. Our research contributes to the ongoing attention to a fundamental question about whether RPTs lead to firm-level efficiencies and value creation and, if so, in what ways. The present study furthers our understanding of whether and how RPT affects firm-level resource allocation efficiencies and firm performance. This study proposes a channel through which RPT affects firm performance. In other words, RPT could affect investment and external financing efficiency, eventually affecting firm performance. Second, we add the growing RPT literature. We provide evidence of negative consequences of certain types of RPTs. The evidence that RPTs are associated with less than optimal investment and that they also appear to be related to the use of internal rather than external financing suggests some concern by external parties about the nature of the RPTs or the potential signal they send about management opportunism. The paper is organized as follows. The next section provides a background on RPT disclosure and discusses prior RPT research, leading to the development of our hypotheses in the following section. The research design we use is presented in the next section. We then discuss our sample, which is followed by a discussion of our results. We then conclude and discuss implications of our research. BACKGROUND INFORMATION AND PRIOR RPT RESEARCH RPT Disclosures Material RPTs are required to be disclosed in the financial statements (FASB ASC 850-10- 50-1). The required disclosures include (1) the nature of the relationship, (2) a description of the transaction, (3) the dollar amounts of the transactions for each income statement period presented, (4) and amounts due to or from related parties at each balance sheet date (FASB ASC 850-10-50-1). This leads to two important questions in understanding the RPT disclosure. 3

First, what is an RPT? An RPT is any transaction (for example, sales, purchases, services, and borrowings) between the company and a related party. FASB defines related parties as an affiliate of the company, investments accounted for under the equity method of accounting (or the fair value election), trusts for the benefits of employees, principle owners of the company, management, and parties with significant influence over management or the operating policies of one of the transacting parties (FASB ASC 850-10-20). Related parties also include the immediate family members of the principle owners and management. Second, why is disclosure required? In promulgating the original accounting standard, the starting point for FASB was that these types of transactions are not considered arm s length (FASB 1982, 3). As such, the potential for the related party to act opportunistically to determinant of other parties, such as unrelated shareholders. Financial statement reliability may also be affected when companies are involved in RPTs (FASB 1982, 15). Financial statement users should therefore be aware of RPTs when making investing and other business decisions. The Securities and Exchange Commission (SEC) is also concerned about RPTs and has separate disclosure requirements. Specifically, companies must disclose certain relationships and related transactions information in their SEC filing, such as registration statements, annual reports, and proxy statements. Disclosure (in the financial statements or other SEC filing) is required for any RPT involving more than $120,000 and in which the related persons had a direct or indirect material interest, naming such person and indicating the person's relationship to the registrant, the nature of such person's interest in the transaction, the amount of such transaction and, where practicable, the amount of such person's interest in the transaction (SEC 2004b, subsection 229.400a). 1 The specific dollar limit may result in disclosure of RPTs in the SEC 1 Prior to December 15, 2016, the disclosure limit was only $60,000. 4

filing that are not considered material for financial statement reporting purposes (and vice versa). Financial statement users therefore need to review multiple sources to fully understand the company s RPTs. In addition, the company s policy and procedures concerning the approval of RPTs is included as part of the SEC s required disclosures for filings after December 15, 2006 (SEC 2006). Prior RPT Research Research on RPTs has expanded in recent years. Much of the research is international in nature as information on RPTs is incorporated in commercial databases of company information. That is not the case in the U.S. where hand collection of the RPT information is necessary. Therefore, RPT research on U.S. companies is more limited. Here, we focus on RPT research related to management behavior. A number of studies investigate the valuation implications of RPTs. Using a U.S. setting, Kohlbeck and Mayhew (2010) find the firms that disclose RPTs have significantly lower valuations and marginally lower subsequent returns. Further, RPT type and party affect these market perceptions. In addition, Ryngaert and Thomas (2012) use a pre-post design and find that RPTs are only detrimental to firm value if the transaction was entered into after the counter party is considered a related party. They suggest these results are consistent with ex post RPTs being opportunistic. Similar negative valuation results are found internationally. Nekhili and Cherif (2011) study French firms and find that RPTs involving directors, officers and major shareholders have a negative impact on firm value. Initially, Ge, at al. (2010) find that Chinese firms with RPTs involving sales of goods or assets have lower valuation multiples. However, the negative valuation effects disappear with new regulations requiring fair value measurement of RPTs. Lo 5

and Wong. (2016) finds that disclosure of the transfer pricing method is incrementally valuerelevant beyond the related party sales amount. A number of studies examine RPTs in the context of earnings management. Chen, et al. (2011) examine RPT-based earnings management during the IPOs of a sample of Chinese firms. They find that RPTs are positively associated with firm s operating performance prior to the IPO; however, the association declines in the post-ipo period. In another study of Chinese firms, Wang and Yuan (2012) find that RPTs appear to have a negative impact on the usefulness of earnings. Research also attempts to explain factors associated with motivations behind RPTs. Propping earnings can be accomplished with RPTs. For example, Jian and Wong (2010) analyze Chinese firms and document the use of abnormal related party sales in order to prop up earnings of listed firms combined with lending back to the related party. Jian and Wong show that using abnormal sales complements accruals management in meeting earnings targets. Williams and Taylor (2013) also find that abnormal sales are used to prop earnings in China. They further find that it is more common among firms at risk of being delisted or with greater state-based ownership. Tunnelling is another RPT mechanism where the related party extracts economic resources for the benefit of the related party. Ahorny, et al. (2010) investigate 185 Chinese IPOs. In addition to providing evidence on the use of RPTs to prop earnings prior to the IPO, Aharny, et al (2010) also provide evidence of opportunistic tunnelling after the IPO. Tunnelling took the form of non-repayment of related party loans. Jia et al. (2013) investigate transfers of resources among related parties within a business group in China. They find that loans are used to provide funding to a related party experience credit problems. Likewise, non-loan RPTs are used to support related parties experiencing lower firm performance. 6

Finally, researchers consider the effect of governance and regulation on RPTs. For example, Bennouri, et al. (2015) investigate whether entering into RPTs is mitigated when the firm has a higher quality auditor. Investigating 85 French firms, Bennouri, et al. (2015) find Big 4 clients report fewer RPTs. These results are consistent with auditors being concerned about negative RPT implications for their reputation and discouraging management from entering into RPTs. Balsam, et al. (2017) find that RPTs are more likely when the firm has weaker corporate governance and when efficient contracting motivations exist. Further, the efficient contracting motivation results are limited to the period after the 2006 SEC disclosure change. Hwang and Chiou (2013) find that the effect of changes in disclosure regulation in Taiwan mitigates earnings management associated with RPTs with Chinese entities In summary, competing arguments exist that RPTs represent either opportunistic behavior or efficient contracting. However, the empirical evidence to date tends to support the opportunistic argument. Studies mostly find RPTs are value decreasing, but any decrease is affected by the type of RPT as well as the classification of the related party. RPTs are also associated with earnings management, propping, and tunneling. Negative effects are mitigated to a certain degree by stronger governance. HYPOTHESES Our primary hypotheses concern the association of RPTs and capital allocation. We acknowledge that all RPTs are not alike. Therefore, associations with capital allocation may vary based on the nature of RPTs. We first provide a discussion of the two competing arguments behind RPTs that supports that RPTs are not alike. We then develop our specific capital allocation hypotheses as it relates to RPTs. 7

Prior research generally identifies two reasons why firms enter into RPTs. On the one hand, agency theory suggests the possibility that RPTs are opportunistic and that managers will over consume perquisites. RPTs that favor the related party to the firm s detriment represent examples of perquisite consumption (i.e. inappropriate wealth transfers). This over-consumption damages the firm s stakeholders (Jensen and Meckling 1976). Consistent with this opportunistic behavior, the FASB in its basis of opinion for the RPT disclosure standard was also concerned that related parties involved in transactions with the firm would put themselves ahead of the firms (FASB 1982). On the other hand, contracting theory suggests that RPTs are a form of efficient contracting. RPTs provide an efficient approach to structuring a transaction as there is less information between the parties. As a substitute for other cash-based compensation, or a more liquid form of compensation compared to stock option, RPTs can be used as part of a compensation package. RPTs therefore may meet the needs of the company (Kohlbeck and Mayhew 2010). Consistent with this notion, no country has completely banned RPTs (Djankov et al 2008). The above discussion suggest that RPTs may be good (form of efficient contacting) or bad (opportunistic behavior). We refer to RPTs associated with these two motivations as business and tone, respectively. It is likely that RPTs of both types exist. Indeed, Kohlbeck and Mayhew (2010, 2017) find evidence supporting both competing arguments. This differential impact is incorporated in our hypotheses on capital allocation. As discussed above, managers may use certain RPTs, identified as tone RPTs, opportunistically. Specifically, managers may use tone RPTs to advance their careers or other personal agendas. Tone RPTs may also be a way for management to entrench themselves by currying favor with other key stakeholders who might help them to resist shareholder discipline. 8

In addition, prior studies (e.g., Balsam, et al 2017) find that RPTs are associated with weaker corporate governance. So, tone RPTs could be used as a self-serving tool for managers to extract private benefits, entrench, shirk, build empires, or consume perks that could cause investment inefficiencies. However, the efficient contracting motivation behind business RPTs suggests that these transactions are in the firm s best interest. In this situation, RPTs would not cause investment inefficiencies. We expect that investment efficiency is not affected by business RPTs any different than other transactions. Based on this argumentation, we pose the following hypothesis: H1: Investment sensitivity to Q is less (no different) for tone (business) RPTs. A firm s internal funds (i.e., internally generated cash flow) are important in determining a firm s level of investment (for example, see Blundell et al., 1992; Whited, 1992; Hubbard and Kashyap 1992). As a result, firms investments are associated with internally generated cash flow. The presence of certain RPTs may alter this relationship. Tone RPTs are likely to increase the association between investment and internally generated cash flow for at least two reasons. First, the agency conflict linked with tone RPTs may result in financiers realizing any distortions in capital allocation and consequently restricting the firm s access to the external financing. Second, entrenched managers of tone RPT firms may rely on internally generated cash for investments rather than face potential disciplinary forces involving access to external capital markets. Combined, investments will be more dependent on internally generated cash in the presence of tone RPTs. 9

However, no such difference is expected for business RPTs. Similar to the discussion for the first hypothesis, business RPTs are in the best interest of the firm and should not affect a firm s access to the external capital markets. We therefore expect that investments of firms reporting tone RPTs are associated with more internally generated cash, while no differential association is expected for business RPTs. Our second hypothesis is stated as follows: H2: Investment sensitivity to internally generated cash is more (no different) for tone (business) RPTs. We adopt a similar framework for our two external financing hypotheses. High marginal Q firms invest more and therefore are required to raise more capital to fund the investment. The external financing Q relationship is therefore positive. However, if tone RPTs distort investment sensitivities to Q as predicted in our first hypothesis, external financing is also likely affected negatively. Tone RPTs are expected to have a detrimental effect on the external finance- Q relationship. Investors may also view tone RPT activities as strategies that maximize managerial wealth and satisfaction at the expense of investors wealth (consistent with the opportunistic nature of tone RPTs). Consequently, external financing could be more dependent on internally generated cash for tone RPT firms. Since business RPTs are assumed to be efficient contracting, we don t expect differential effects on either external financing sensitivity for business RPTs. In line with these arguments we postulate the following two hypotheses: H3: External financing s sensitivity to Q is less (no different) for tone (business) RPTs. 10

H4: External financing s sensitivity to internally generated cash is more (no different) for tone (business) RPTs. Our final hypothesis concerns the effect of any capital allocation inefficiencies associated with RPTs on future firm performance. In our hypotheses discussed above, we expect tone RPTs to be associated with increased investment and external financing inefficiencies, consistent with the opportunistic nature of these transactions. Therefore, tone RPTs should lead to poorer future financial performance. Business RPTs, on the other hand, are not predicted to be associated with any investment or external financing inefficiency. By extension, we would therefore not expect an association between business RPTs and future firm performance. Our fifth hypothesis is stated as follows: H5: Tone (business) RPTs are negatively (not) associated with future financial performance. In summary, we expect that tone RPTs are associated with investment inefficiencies in terms of sensitivities of 1) investments to both Q and internally-generated funds, and 2) external financing to both Q and internally-generated funds. This leads to a negative impact on future firm performance. In contrast, business RPTs are not expected to be associated with investment inefficiencies, external financing inefficiencies, or future firm performance. RESEARCH DESIGN Three different models are required to test our capital allocation hypotheses. Our first model considers a firm s investment sensitivity. We adopt the empirical specifications from Baker et al. (2003), Mclean et al. (2012), and Fazzari et al. (1988). 2 INV = β0 + β1 Qt-1 + β2 CFt-1 + β3 PBVt-1 + Firm FE + Year FE + εt (1) where variables are defined in Appendix 1. 2 Firm and year subscripts for all equations are omitted except to indicate a period other than the current year. 11

The dependent variable (INV) is computed as the sum of the yearly growth in property, plant, and equipment, and R&D expense, divided by prior year total assets (Mclean et al, 2012). For Tobin s Q, we follow Baker et al. (2003). Tobin s Q is estimated as the book value of total assets plus the market value of equity less the book value of equity divided by the book value of total assets. Internally generated cash flows (CF) are computed as net income before extraordinary items plus depreciation and amortization expense and R&D expense, divided by the book value of total assets (Mclean et al. 2012). We also include the industry-adjusted market to book value of equity ratio to control for overvaluation (Baker et al. 2003). The estimated coefficient for Q represents the investment sensitivity to Tobin s Q while the estimated coefficient for CF represents the investment sensitivity to internally generated cash flow. Consistent with prior research (for example, Tobin, 1969; Fazzari et al, 1988), positive coefficients are expected for both. Our control for overvaluation (PBV) is expected to have positive relationship with investment (Baker et al. 2003). 3 Our first two hypotheses are concerned with differential investment sensitivities for firms reporting RPTs. Two RPT classifications are considered tone and business. As discussed above, tone RPTs are considered to be more representative of opportunistic behavior while business RPTs are related to efficient contracting. We modify Equation (1) to include RPT variables as follows: INV = β0 + β1 (Qt-1 x RPTt-1) + β2 (CFt-1 x RPTt-1) β3 Qt-1 + β4 CFt-1 + β5 RPTt-1 + + β6 PBVt-1 + Firm FE + Year FE + εt (2) where RPT is either TONE or BUSINESS, and variables are defined in Appendix 1. 3 Baker et al. (2003) show that firms invest more when their stock is overvalued. 12

Our two RPT variables are TONE and BUSINESS. The two RPT variables are determined consistent with the classifications in Kohlbeck and Mayhew (2017). The classifications are a function of both the nature of the transaction and the related party (see Appendix 2). Tone RPTs (TONE) proxy for opportunistic RPTs and include transactions with directors, executive officers, and major shareholders (DOS) such as loans, borrowings, guarantees, consulting, legal and investing services, unrelated business activities, overhead reimbursements, and stock transactions. Tone RPTs also include unrelated business activities, consulting, and legal and investing services where the counter party is an investment of the company (Investee). On other hand, business RPTs (BUSINESS) are considered a function of efficient contracts. These transactions include Investee loans, Investee borrowings, Investee guarantees, DOS and Investee related business activities, DOS and Investee leasing activities, Investee overhead reimbursement, and Investee stock transactions. In our first hypothesis, we expect that tone RPTs distort the investment sensitivity to Q (β1 < 0) and no differential effect is expected for business RPTs (β1 = 0). The second hypothesis predicts that the investment sensitivity to internally generated cash flows will increases with tone RPTs (β2 > 0) and is not affected by business RPTs (β2 = 0). Ex ante, we are unable to make any predictions for the main effects of either RPT variable. The third and fourth hypotheses consider external financing s sensitivity to Q and internally generated cash flow. Similar to prior research (for example, Baker et al. 2003; Mclean et al. 2012), we modify Equation (2) to include external financing as the dependent variable. EXTFINt = β0 + β1 (Qt-1 x RPTt-1) + β2 (CFt-1 x RPTt-1) + β3 Qt-1 + β4 CFt-1 + β5 RPTt-1 + β6 PBVt-1 + Firm FE + Year FE + εt (3) where RPT is either TONE or BUSINESS, and variables are defined in Appendix 1. 13

We consider both equity and debt as sources of external financing. External equity finance is estimated as the change in book value of equity plus change in deferred income taxes less the change in retained earnings, divided by prior period total assets. External debt financing is computed as the sum of changes in debt included in current liabilities and long-term debt, divided by prior period total assets. The two are then combined to form our dependent variable (EXTFIN). Our predictions for the control variables are consistent with prior research (for example, see Mclean et al. 2012, among others). We expect that Tobin s Q is positively associated with external financing as the firm faces less financial constraints. As for internally generated cash, we expect an inverse relationship because lower cash flow firms are more likely to require external funds. In our third hypothesis, we expect that the external financing sensitivity to Q is less for firms reporting tone RPTs (β1 < 0) and no different for those reporting business RPTs (β1 = 0). For the fourth hypothesis, the investment sensitivity to internally generated cash flows is expected to increase with tone RPTs (β2 > 0) and not be affected by business RPTs (β2 = 0). Our final hypothesis is concerned with the association of tone and business RPTs with future financial performance. We consider two forms of financial performance accounting and share price - measured over the three year period from year t+1 to t+3. We use industry fixed effects (using Fama-French 48 industry classification) and year fixed effects in our firm performance models. 4 4 Using firm fixed effects identification in our firm performance empirical framework has its own limitation. Firm fixed effects identification requires time series variations in our main firm performance measures. But these variables do not vary much from year to year to fully support the firm fixed effects model because our sample firms 14

Consistent with Chen et al. (2007) and Bhandari and Javakhadze (2017), we use the following models for accounting and stock-based performance to test H5. PERFACCt+3 = α1 + α2 RPT+ α3 KZ4 + α4 SALESt-1 + α5 Q + α6 LNMV + Industry FE + Year FE + εt (4) PERFSTOCKt+3 = α1 + α2 RPT + α3 LNMV + α4 BMt-1 + α5 DEBTAT + α6 RETMOM + Industry FE + Year FE + εt (5) where RPT is either TONE or BUSINESS, and variables are defined in Appendix 1. Future accounting performance (PERFACC) is computed as either the average of the ratio of earnings before interest, taxes, depreciation, and amortization to the sum of the market value of equity plus total debt for the three years ending t+3, or the average annual sales growth over the same period. Equation (4) also controls for other factors that may affect future operating performance including capital constraints (measured as the four variable version of Kaplan and Zingales (1997) index of capital constraints), Tobin s Q (Q), and firm size (SALES and LNMV). We expect that accounting performance is increasing in Tobin s Q and firm size, and decreasing in capital constraints. For Equation (5), the dependent variable, future share price performance is captured by the natural log of one plus alpha. Alpha is estimated over the three year period ending in year t+3 using either the Fama French 3-factor model or the Carhart 4-factor model. We also control for other variables that may affect future share price performing including include size (LNMV), the book-to-market ratio for potential undervaluation (BM), firm leverage (DEBTAT), and stock are from the S&P 1500 firms and there is insufficient within-subject variation in our firm performance measures. Hence we use industry fixed effects. 15

momentum over the previous 24 month period (RETMOM). We expect that share price performance is increasing (decreasing) in the book-to-market rate and firm leverage (stock momentum). We make no prediction regarding firm size. For both future performance metrics, our fifth hypothesis predicts a negative coefficient for tone RPTs (α2 < 0), future performance will be lower. As for business RPTs, no association with future firm performance is expected. SAMPLE We start with the S&P 1500 firms in 2001, 2004, 2007, 2010 and 2013. We identify 6,489 firm-year observations from 2001, 2004, 2007, 2010, and 2013 for which complete financial statement information is available to identify whether or not the firm reported RPTs. We eliminate 192 firm-year observations missing asset values, amounts to calculate independent variables, and prior year observations needed to calculate lagged variables. Our final sample for the capital allocation tests consists of 6,297 firm-years. An additional 953 firm-year observations without sufficient data to compute future performance measures are then eliminated. The future firm performance sample consists of 5,344 firm-year observations (see Table 1, Panel A). Table 1, Panel B reports distributions of RPT firms by year for the capital allocation sample. Approximately 66 percent of the sample reports RPTs in 2001. This drops over the subsequent years to 47 percent in 2013. The decrease is consistent with the ban of related party loans and increased scrutiny of RPTs in general. A closer look at the data indicates that business RPTs are relatively flat over time. However, Tone RPTs decline from 54% in 2001 to 22% in 2013. 16

Table 1, Panel C reports distributions of the capital allocation sample by Ftama-French 12 industry classifications. Firms in financial industry have the largest RPT percentage overall and separately for tone and business RPTs. Manufacturing firms have the lowest level of RPTs. Descriptive statistics for variables included in Equations (2) and (3) are reported in Table 2, Panel A, using the capital allocation sample. Overall, approximately 18% of the sample firms report only tone RPTs, 19% report only business RPTs, and 18% report both tone and business RPTs. Investments averages 7.3% of prior period total assets while external financing averages 3.2% of prior period total assets. However, a substantial number of the firms have negative investments or equity financing. The mean (median) Tobin s Q is 1.9 (1.5), consistent with prior research. The cash flow for the middle 50% of firms range from 5.6% to 17.5% of prior period total assets. Table 2, Panel B, present the descriptive statistics for the future firm performance sample. RPT data is similar to that reported in Panel A. The mean (median) annual three-year return on assets and sales growth are 9.6% (9.8%) and 7.5% (6.2%), respectively. Both the mean 3- and 4- factor alphas suggest positive shareholder returns. The mean book-to-market ratio is 0.48 (median is 0.43) while the debt to total assets is 19% (median of 18%). Pearson correlation matrices among the equation variables for both samples are reported in Table 3. By construction, both tone and business RPTs are highly correlated with RPT in both samples; however, these variables are not included in any estimation at the same time. Among the capital allocation variables (Panel A), the correlations between all of the RPT variables and either investments or external finance are not significant. Significant correlations in the future firm performance sample (Panel B) between tone or business RPTs and future accounting performance (future share performance) are positive (not significant). This univariate evidence 17

does not support our hypotheses. The only significant correlations above 0.50 among the control variables in either sample are between 1) Tobin s Q and both cash flows and industry-adjusted price to book value ratio, 2) KZ4 and the ratio of debt to total assets, and 3) sales and the natural log of market value. 5 None of these correlations are expected to affect our analysis. RESULTS RPTs and Capital Allocation We first analyze the association between RPTs and both investment-q and investment-cash flow sensitivities. We estimate Equation (2) separately for tone and business RPTs and report our results in Table 4, Panel A. Standard errors are corrected for heteroscedasticity and clustered at the firm level. Fixed effects are included for year to capture factors specific to individual years and firm to account for unobserved, time-invariant firm-level heterogeneity. 6 The adjusted R 2 s of approximately 28% are consistent with prior research (e.g., Bhandari and Javakhadze 2017). We control for the direct effects of a particular RPT type, Q, and CF on investments to limit the potential that these direct effects drive our results. The estimated coefficients for Q and CF are both positive and significant in both regressions, consistent with our predictions. The other two control variables are not significant. Our variables of interest to test the first and second hypotheses are the estimated coefficients for β1 and β2 on the interaction terms. 5 The 3- and 4-factor alphas are also highly correlated, but are different constructs for future shareholder returns and their correlation is expected. 6 Industry fixed effects are irrelevant in a firm fixed effects framework and are not included in our regressions with firm fixed effects. 18

In the first column, RPTs are represented by tone RPTs. The coefficient estimate for Q*TONE is significantly negative (-0.0253, p-value < 0.01) suggesting that the relation between Q and investment is weaker for firms reporting tone RPTs. In addition, the coefficient estimate for CF*TONE is significantly positive (0.2365, p-value < 0.05) implying that the effects of internally generated cash flow on investments is stronger for firms reporting tone RPTs. These results are consistent with H1 and H2 that tone RPTs distort investment sensitivity to Q as well as increases a firm s investment dependency on internally generated cash. The tone RPT effects are also economically significant. The estimation using business RPTs is in the second column. Both interactions, Q*BUSINESS and CF*BUSINESS, are not significant. These results are also consistent with our first and second hypotheses that business RPTs do not distort the investment sensitivity to Q or affect a firm s investment dependency on internally generated cash. We then include the natural log of total assets as an additional control variable to control for the effect of firm size on investments and report our results in Table 4, Panel B. The estimated coefficient for LNAT is negative and significant in both regressions suggesting that investments levels are influenced by firm size. In addition, the direct effect for tone RPTs is now positive and significant indicating the investments is greater when firms report tone RPTs. As for the interaction terms and the tests of our hypotheses, our results are similar to those reported in Panel A. We also consider an alternative measure of investments where we define it as the year-toyear growth in total assets. Our results are qualitatively similar to our earlier findings with the exception that the estimated coefficient for CF*TONE continues to be positive but is no longer significant. 19

Next we examine the associations between RPTs and the external financing-q and external financing-cash flow sensitivities. We estimate Equation (3) explaining external financing separately for tone and business RPTs and report our results in Table 5, Panel A. 7 Significant coefficient estimates for the control variables are largely consistent with prior theoretical and empirical evidence (e.g., Mclean et al 2012). The interaction terms are used to test our third and fourth hypotheses. When tone RPTs is used in the first column, we find a negative estimated coefficient for the Q interaction term (- 0.0136, p-value < 0.05) suggesting that the sensitivities of external financing to Q is weaker for firms reporting tone RPTs. This result is consistent with our theoretical predictions and H3. With respect of cash flow sensitivities, the coefficient estimate for CF*TONE is significantly positive (0.1083, p-value < 0.10). This result implies that the sensitivity of external financing to cash flow is stronger for firms reporting tone RPTs. In Table 5, Panel B, we re-estimate Equation (3) with the addition of the natural log of total assets to control for the effect of firm size on external financing. The estimated coefficient for LNAT is negative and significant in both regressions suggesting that external financing is influenced by firm size. The estimated coefficients for the interaction terms and the tests of our hypotheses are similar to those reported in Panel A with the exception that CF*TONE is positive, but is only marginally significant with a one-tailed test. Next, we use an alternative measure of external financing from Kisgen (2006). External financing is defined as the sum of equity financing (measured as the sale of common and 7 Standard errors are corrected for heteroscedasticity and clustered at the firm level. Fixed effects are included for year to capture factors specific to individual years and firm to account for unobserved, time-invariant firm-level heterogeneity. 20

preferred stock less the purchase of common and preferred stock from year t to t+1) and debt financing (measured as long-term debt issuance less long term debt reduction plus changes in the current debt from year t to t+1), all divided by prior year total assets. We re-estimate both models and our results are qualitatively similar to those reported earlier with the exception that the estimated coefficient for CF*TONE continues to be positive but is no longer significant. In the aggregate, our findings provide strong evidence that tone RPTs distort the investment and external financing sensitivities to Q. In addition, cash flow is more positively associated with investment and to some extent to external financing for firms reporting tone RPTs. Consistent with our expectations, our analysis provides no evidence of an association between business RPTs and either investment or external financing sensitivities to Q and cash flows. RPTs and Future Firm Performance We next examine the associations of tone and business RPTs with future firm performance. First, we analyze future firm performance in terms of accounting numbers and estimate Equation (4) separately for tone and business RPTs. 8 In Table 6, Panel A, future accounting performance is defined as average return on assets over the future three year period. The explanatory power for both estimations are approximately 12%. The estimated coefficients for the control variables are also similar between the two estimations. Return on assets are increasing in sales and market value of equity, but decreasing in Tobin s Q. As for our hypothesis tests, the estimated coefficient tone RPT is negative and significant (-0.0029; p-value < 0.10) suggesting tone RPTs are associated with lower future 8 Fixed firm- and year-effects are included in each estimation and standard errors are corrected for heteroscedasticity. 21

return on assets. The estimated coefficient for business RPTs is not significant. Our fifth hypothesis is supported when defining future accounting performance as return on assets. We next define future accounting performance as sales growth and report our results in Table 6, Panel B. The explanatory power of the two models are approximately 16%. The estimated coefficients for all four control variables are significant in both estimations. Sales growth is increasing (decreasing) in KZ4, Tobin s Q, and market value of equity (sales). However, the estimated coefficient for tone (business) RPT is not significant (positive). These results are not as expected but are the differential impact between the two estimations is consistent with our hypothesis higher sales growth for business RPTs compared to tone RPTs. In untabulated analyses, we also consider ROA and sales growth over the subsequent five years. Our sample size is reduced to 3,875 firm-year observations with the additional data requirements. Our results weaken, but are qualitatively similar to those presented. Second, we define future firm performance in terms of stock-based performance and estimate Equation (5) separately for tone and business RPTs. 9 We use two measures of stock-based performance where we vary measuring alpha using a 3-factor versus a 4-factor model. Table 7, Panel A reports the results of estimating Equation (5) using the alpha based on the 3- factor model. The explanation power is 2.3%, consistent with prior research. The only control variable with a significant coefficient is the natural log of market value of equity which is inversely associated with the 3-factor alpha. With respect to our RPT variables, the estimated coefficient for tone RPTs is negative and significant (-0.0347; p-value < 0.05) and that for business RPTs is not. Our fifth hypothesis is supported. 9 Fixed industry- and year-effects are included in each estimation and standard errors are corrected for heteroscedasticity. 22

In Table 7, Panel B, we repeat the analysis with the 4-factor alpha as our dependent variable. The explanatory power is just under 1% and none of the estimated coefficients for control variables are significant. However, the estimated coefficient for tone RPT is again negative and significant (-0.0318; p-value < 0.10) and the estimated coefficient for business RPTs is not, supporting our fifth hypothesis. We also consider alternative measures and specifications to test robustness of the stock performance results (untabulated). Instead of three years, we measure the 3-factor and 4-factor alphas over five years. We measure stock performance as the Sharpe ratio, estimated as the ratio of stock excess return over the three-year and five-year periods divided by the standard deviation of stock returns for the same period. Our results are qualitatively similar to those reported earlier. Finally, we establish a link between specific type of RPT, investment efficiency, and performance following methodology developed by prior research (e.g., Jiraporn and Liu 2008). More specifically, we estimate two investment regressions - Equations (1) and (2) where the difference is the inclusion of the RPT variables. The difference in the predicted values of investment from these two regressions is the change in investments directly attributed to the RPT type (tone vs. business) included in Equation (2). Next, we use this change in investments (denoted ΔINVTONE and ΔINVBUS) as the main independent variable in Equation (5), instead of RPT type. If tone RPT distorts investment efficiency, it should be reflected in firm performance. The untabulated analysis shows that the estimated coefficient for ΔINVTONE is negative and significant indicating that tone RPT distorts investment efficiency which is detrimental for firm performance. The estimated coefficient for ΔINVBUS is not significant. These results are consistent with the expectations of the fifth hypothesis. 23

CONCLUDING THOUGHTS We investigate whether RPTs affect firm-level resource allocation efficiency. More specifically, we test whether RPTs classified as tone and business RPTs are associated with investments sensitivity to both Q and cash flow. Consistent with the agency theory view, we posit and find evidence consistent with opportunistic RPTs (tone RPTs) redirecting amounts to related parties rather than investments. We further expect and find that investment is more dependent on internally generated cash for firms with these higher opportunistic RPT activities. We posit two reasons that could justify the positive effect on the sensitivity of investment to cash flow. First, because of agency considerations, financiers could realize distortions in capital allocation; consequently, restrict access to the external financing. Thus, investments will be more dependent on internally generated cash. Second, entrenched managers of opportunistic RPT firms could try to shield themselves from disciplinary forces of external capital markets by relying on internally generated cash for investments. High marginal Q firms should raise more capital as they invest more. However, distortions in investment sensitivities to Q caused by opportunistic RPTs can be manifested in external financing. If RPT activities result in investment inefficiencies, then opportunistic RPTs should have a detrimental effect on the external finance-q relationship. In addition, investors may view opportunistic RPT activities as strategies that maximize managerial wealth and satisfaction at the expense of investors wealth. Consequently, we predict and provide evidence that external financing is less dependent on current growth opportunities (Tobin s Q) and more dependent on internally generated cash for these firms. 24

Finally, we consider whether or not these inefficiencies in capital allocation are manifested in lower future firm performance. We test whether or not tone and business RPTs are associated with future accounting and stock performance. We find that future firm performance is lower for tone RPT firms but not business RPTs, consistent with economic impacts of the capital allocation inefficiencies. Combined, we provide evidence consistent with negative effects associated with tone RPTs, but not business RPTs. Our evidence is consistent with the opportunistic nature of tone RPTs and the efficient contracting view of business RPTs. 25

References Aharony, J., J. Wang, H. Yuan. 2010. Tunneling as an incentive for earnings management during the IPO process in China. Journal of Accounting and Public Policy 29 (1): 1-26. Baker, M., J. Stein, and J. Wurgler. 2003. When does the market matter? Stock prices and the investment of equity-dependent firms. Quarterly Journal of Economics 118 (3): 969 1005. Balsam, S., R. Gifford, and J. Puthenpurackal. 2017. Related party transactions, corporate governance, and CEO compensation. Journal of Business, Finance and Accounting 44 (5): 854-894. Bennouri, M., M. Nekhili, and P. Touron. 2015. Does auditor reputation discourage relatedparty transactions? The French case. Auditing: A Journal of Practice & Theory 34 (4): 1-32. Bhandari, A., and D. Javakhadze. 2017. Corporate social responsibility and capital allocation efficiency. Journal of Corporate Finance 43: 354-377. Blundell, R., S. Bond, M. Devereux, and F. Schiantarelli. 1992. Investment and Tobin s Q: Evidence from company panel data. Journal of Econometrics 51: 233 57 Chen, Q., I. Goldstein, W. Jiang. 2007. Price informativeness and investment sensitivity to stock price Review of Financial Studies 20: 619 650. Chen, J., P. Cheng, and X. Xiao. 2011. Related party transactions as a source of earnings management. Applied Financial Economics 21 (3): 165-181. Djankov, S., R. La Porta, F. Lopez-de-Silanes, and A. Shleifer. 2008. The law and economics of self-dealing. Journal of Financial Economics 88, 430-465. Fazzari S., R. Hubbard, and B. Petersen. 1988. Financing constraints and corporate investment, Brookings Papers on Economic Activity 1, 141 206. 26

Financial Accounting Standards Board Accounting Standard Codification (FASB ASC) available at http://www2.aaahq.org/asclogin.cfm. Financial Accounting Standards Board. 1982. Statement of Financial Accounting Standards No. 57, Related Party Disclosures. Norwalk, CT: FASB (FAS 57). Ge, W., D. Drury, S. Fortin, F. Liu, and D. Tsang. 2010. Value relevance of disclosed related party transactions. Advances in Accounting 26 (1): 134-141. Hubbard, R., and A. Kashyap. 1992. Internal net worth and the investment process: An application to U.S. agriculture. The Journal of Political Economy 100 (3): 506-534. Hwang, N., J. Chieh, and Y. Wang. 2013. Effect of disclosure regulation on earnings management through related-party transactions: Evidence from Taiwanese firms operating in China. Journal of Accounting and Public Policy 32 (4): 292-313. Jensen, M. and W. Meckling. 1976. Theory of the firm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3 (October): 305-360. Jia, N., J. Shi, and Y. Wang 2013. Coinsurance within business groups: Evidence from related party transactions in an emerging market. Management Science 59 (10): 2295-2313. Jian, M., and Wong, T. 2010. Propping through related party transactions. Review of Accounting Studies 15 (1): 70-105. Jiraporn, P., & Liu, Y. (2008). Capital structure, staggered boards, and firm value. Financial Analysts Journal 64 (1): 49-60. Kaplan, S., and L. Zingales. 1997. Do investment-cash flow sensitivities provide useful measures of financing constraints, Quarterly Journal of Economics 112: 169 215. Kisgen, D. J. (2006). Credit ratings and capital structure. The Journal of Finance 61 (3): 1035-1072. 27