Printer Friendly

Does earnings management relieve the negative effects of mandatory pension contributions?

Mandatory pension contributions (MCs) are negative shocks to a firm's liquidity that can unfavorably impact its cost of capital, financing, and investment plans. We examine whether firms faced with MCs use both noncash (NEM) and cash-generating earnings management (CEM) to partially offset their negative effects. Firms increase CEM, but not NEM, when they experience MCs. We also find that earnings management associated with MCs does not substantially lower the weighted cost of capital or boost external funding and investment. Our findings suggest that MC firms use CEM as it directly generates cash to fund MCs, while NEM does not.

**********

Mandatory pension contributions (MCs) are often unexpected negative shocks to a firm's liquidity that can raise its cost of capital (Campbell, Dhaliwal, and Schwartz, 2012) and lead to reduced investment (Rauh, 2006). Linck, Netter, and Shu (2013) find that financially constrained firms can use earnings management to signal profitable investment opportunities and help attract investor funds for those investments. Our paper examines whether firms faced with MC shocks increase their earnings management to partially offset the rise in the cost of capital and the fall in investment associated with MCs.

The accounting literature establishes two general forms of earnings management: 1) accrual-based earnings management and 2) real earnings management. The more widely studied accrual-based methods have little or no impact on cash. For example, if a firm reduces its estimate of uncollectable accounts, reported earnings increase, but not cash flow. Conversely, real earnings management impacts cash. For example, a firm could expand production to spread its fixed costs and boost profit margins and earnings, even though this method of real earnings management consumes cash.

In our study, we group the various methods of earnings management into two different categories: 1) cash-generating earnings management (CEM) and 2) noncash-generating earnings management (NEM). (1) We do this as we wish to study how managers cope with the negative cash shocks caused by MCs. CEM can generate internal cash to relieve financing constraints directly, but can also do so indirectly by boosting reported earnings to attract more investors and external financing. NEM relies solely on the indirect channel of boosting earnings to attract external financing.

Whether MC firms use CEM, NEM, some combination, or neither, is an empirical issue and depends upon their relative costs and effectiveness in boosting reported earnings and attracting financing. In addition to generating cash directly, CEM has the advantage that it is more difficult for investors and auditors to detect than NEM as changes in business operations are not required to be disclosed. Indeed, Graham, Harvey, and Rajgopal (2005) and Cohen, Dey, and Lys (2008) find that managers prefer it to NEM, especially after Congress passed the Sarbanes-Oxley Act (SOX) in 2002. SOX requires managers to certify the integrity of their financial statements, but not the soundness of their operating decisions.

However, both CEM and NEM have drawbacks. CEM can be costlier to implement than simple accounting book adjustments as it involves changing, and perhaps disrupting, a firm's physical operations. The physical constraints of firm production could also make CEM a less flexible or potent method to boost earnings than NEM. A corresponding drawback of NEM is that investors must recognize the earnings signal and perceive it as a signal of good investment opportunity as opposed to deception.

Defined benefit (DB) plans are substantial liabilities for US firms. During our sample period from 1991 to 2013, the average DB plan's assets represent 18.20% of firms' book assets and 36.49% of their market value of equity. In a DB plan, the sponsoring firm promises to make fixed monthly pension payments to its retirees. When the present value of those payments exceeds the DB plan's asset value, the plan becomes underfunded and the sponsoring firm is required to make MCs. Business media reports that the recent drop in interest rates has caused the value of US corporate DB plans to reach their most underfunded levels on record. Although DB plans may be costly and risky, they could be valuable if they attract and retain talented employees.

Rauh (2006) suggests that MCs are substantial enough to adversely affect a firm's investments and competitive position. He finds that, on average, a firm's capital expenditures decrease by $0.60 to $0.70 per dollar of MCs and that industry competitors take advantage by increasing their investments. Franzoni (2009) argues that MCs lead to subsequent negative stock returns, while Campbell et al. (2012) report that MCs increase firms' cost of debt, cost of equity, and weighted average cost of capital (WACC). These studies imply that MCs can impose financial constraints on firms with underfunded DB plans.

Linck et al. (2013) argue that a financially constrained firm with valuable investment projects can manage earnings to help attract additional external capital for its investments. Francis, Olsson, and Schipper (2004) find that smoothed earnings help firms reduce the implied costs of equity. For these reasons, MC firms may be more willing to engage in costly earnings management if it lessens the negative effects of MCs on their financing and investment plans. (2)

Our sample includes only firms with DB plans, some of which experience cash flow shocks from MCs and others that do not. Rauh (2006) and Campbell et al. (2012) find that MCs are driven by the arbitrary and nonlinear structure of pension contribution rules that cause unexpected MC jumps that are unlikely to be endogenous to firms' investment prospects, costs of capital, or external funding opportunities. Specifically, Rauh (2006) argues: "... [t]he function that relates funding status to investment opportunities does not have precisely the same kinks, jumps, and asymmetries as the function that relates pension funding status to required pension contributions." (2006, 34-35). In our case, the function that relates pension funding status to MCs does not have the same exact kinks, jumps, and asymmetries as the function that relates funding status to NEM or CEM. Therefore, MCs provide a relatively clean way to identify whether earnings management helps to counteract the burdens imposed by liquidity shocks, such as MCs. Those burdens can include higher costs of capital, less available external funding, and consequently, less investment.

We begin our analysis by examining the relation between earnings management and MCs while controlling for DB plan funded status and other factors that could explain earnings management. Based on a panel of 1,283 unique firms that sponsor DB plans from 1991 to 2013, we find that firms that are required to make MCs resort to CEM, but not NEM.

Having established that MC firms resort to managing earnings, we next examine whether this earnings management impacts the cost of capital, external financing, and investment. We use MCs interacted with various measures of NEM and CEM as explanatory variables in the cost of capital, external financing, and investment regressions to measure the effects of earnings management conditioned on MCs. Our analysis indicates that earnings management associated with MCs increases the cost of debt, although it can marginally reduce the cost of equity and WACC. In addition, we do not find evidence that earnings management associated with MCs helps firms raise more debt or equity. Intuitively, external creditors would be concerned that any new debt proceeds would go to fund MCs rather than to finance profitable investment projects. Even if the larger managed earnings signal the firm's attractive investment opportunities (Linck et ah, 2013), creditors are less likely than stockholders to benefit from them as their payments are fixed and the new investment returns are uncertain. Consequently, creditors raise their required rate of return on new debt and refuse to lend more to these firms.

CEM associated with MCs may be viewed positively by equity holders as it can generate cash to fund MCs without diluting their holdings. While like Campbell et al. (2012), we find that MCs raise the costs of debt, equity, and WACC, on average, we also determine that CEM can mitigate a small amount of the increase in the cost of equity and overall WACC. However, we find no evidence that CEM increases investment, and some evidence that it actually reduces investment. We also find that MC firms that employ CEM contribute more to their DB plans. Overall, our results can be interpreted as indicating that firms faced with MCs tend to engage in more CEM, but that does not help increase external funding for their profitable investments. Rather, CEM generates internal cash to fund their MCs, and that this is likely to placate their equity holders to some degree.

Our paper contributes to both the pension literature and the earnings management literature. It adds to the pension literature by examining whether firms cope with sudden MCs associated with their underfunded DB pension plans by increasing their earnings management. In addition, it studies whether this earnings management favorably impacts their cost of capital, corporate financing, and investment. Our evidence indicates that firms respond to MCs by managing earnings. We contribute to the earnings management literature by using MCs as exogenous shocks to a firm's liquidity to better identify the potential benefits of earnings management for financially constrained firms. Prior earnings management studies use conventional measures of financial constraints based on firm-level variables, such as size, that can be endogenous to corporate decisions and. as such, may provide biased estimates of the potential benefits (Duchin, Ozbas, and Sensoy, 2010; Hadlock and Pierce, 2010). The only significant benefit of earnings management associated with MCs that we find is that CEM provides cash that firms can use to help fund MCs.

The rest of the paper proceeds as follows. Section I describes the major variable construction and empirical models. Section II provides the data description, empirical results, and discussions. We provide robustness checks in Section III, while Section IV presents our conclusions.

I. The Empirical Models and Measures of Earnings Management and Pension Contributions

A. Measuring Mandatory Pension Contributions and Funded Status

During our sample period, DB firms were required to contribute the larger of the minimum funding contribution (MFC) or the deficit reduction contribution (DRC). MFCs were first stipulated by the Employee Retirement Income Security Act (ERISA) of 1974 that requires sponsors of underfunded DB plans to annually contribute the normal cost, which is the present value of the pension benefits accrued during the year, and installment payments on any unfunded liabilities. DB plan sponsors may amortize the unfunded liability over 5 to 30 years.

The Pension Protection Act of 1987 (PPA 1987) introduced changes that require better funding of DB plans. Specifically, PPA 1987 required that firms deposit 13.75% to 30% of any underfunding into the DB plan as deficit reduction. The remainder of the shortfall can be amortized over three to five years. The Retirement Protection Act of 1994 (RPA 1994) changed funding requirements for years 1995 and later by exempting plans that are over 90% funded from DRCs. RPA 1994 also exempted certain plans that are between 80% and 90% funded, applied the 30% DRC rate to more plans, and raised the lowest DRC rate from 13.75% to 18%.

Similar to Munnell and Soto (2004), Rauh (2006), Bakke and Whited (2012), and Phan and Hegde (2013), we estimate MFC as the normal cost plus 10% of the previous period's funding gap. (3) DRC as a fraction of the funding gap is calculated as the min (0.30, [0.30-0.25 (pension assets/pension liabilities -0.35)]) of the underfunding amount for 1991 to 1994, and as the min (0.30, [0.30-0.40 (pension assets/pension liabilities -0.60)]) of the underfunding amount for 1995 and later. The MC is the larger of the MFC or the DRC. These formulas are nonlinear and can generate jumps in MCs, for example, if the value of the pension assets drop suddenly and the plan becomes substantially underfunded. We calculate MCs at the DB plan level and then aggregate to firm level. Funded status is measured as pension assets minus pension liabilities at the plan level, then aggregated from plan to firm level and divided by the market value of equity at the beginning of the year. (4)

B. The Relations between Earnings Management and MCs

Although earnings management can be costly (Trueman and Titman, 1988; Chaney and Lewis, 1995; DuCharme, Malatesta, and Sefeik, 2004), earnings management can allow financially constrained firms to invest in profitable projects whose benefits outweigh the costs of earnings management. Since MC tightens financial constraints, we expect earnings management to be positively related to MCs.

The MCs associated with underfunded DB plans drain firms' limited internal resources that would otherwise be invested in positive net present value (NPV) projects (Rauh, 2006). To maintain an optimal investment path in the presence of MCs, firms could manage earnings to signal their positive investment prospects and attract external funding. Furthermore, CEM methods of earnings management can generate internal cash. Whether firms use NEM or CEM depends upon the relative costs and benefits of each method of earnings management. CEM could generate more internal funds than NEM and be more difficult to detect, but it could be less effective at boosting earnings to attract external funds and it could be costlier if it disrupts the firm's real operations. Whether firms decide to use NEM or CEM in response to MCs or some combination is an empirical issue.

Our first task is to test for a significant relation between earnings management and MCs after controlling for variables used by earlier studies to explain earnings management. To do this, we use the following model:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (1)

The dependent variable in Equation (1) is either NEM or CEM. Our test variable is MC, and we expect its associated estimate, [[beta].sub.1], to be positive. That is, larger MCs induce greater earnings management. In addition, we control for DB plan funded status, firm size proxied by the natural logarithm of book assets, market-to-book ratio, financial leverage, the natural logarithm of firm age, and volatility of sales growth.

To ensure the robustness of the results, we employ several measures of earnings management used in earlier research, such as Cohen and Zarowin (2010). We include three measures of NEM and three measures of CEM. The details regarding the construction of the measures are provided in Appendix A. The first two measures of NEM depend upon two approaches to measuring performance-matched discretionary accruals. Discretionary accruals, DA1, for firm i at time t are estimated as the difference between the total accruals ([TA.sub.i,t]) and the predicted nondiscretionary accruals ([NDA1.sub.i,t]):

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (2)

We adjust [DA1.sub.i,t] by the contemporaneous [DA1.sub.m,t] of a matched firm m, which is in the same industry as the sample firm and has return on assets (ROA) closest to that of the sample firm. We label the performance-matched discretionary accruals [PDA1.sub.i,t].

[PDA1.sub.i,t] = [DA1.sub.i,t] - [DA1.sub.m,t]. (3)

The second measure of performance-matched discretionary accruals, [PDA2.sub.i,t], is computed in the same way as [PDA1.sub.i,t] except that the "modified Jones" model is used to estimate nondiscretionary accruals for Equation (1). The two methods used to measure nondiscretionary accruals are described in Appendix A.

The last measure of NEM is abnormal production (ABN_PROD). This is where managers abnormally increase production (PROD) to allocate more overhead to inventory and less to costs of goods sold, thereby increasing operating margins and earnings (Roychowdhury, 2006). The computation of ABN_PRODu is the actual firm production costs minus the predicted production costs for firm i at time t. This method consumes cash, perhaps tightening any financial constraint caused by MCs.

Our three measures of CEM are associated with an increase in cash, although not necessarily efficient operations. First, abnormal discretionary expense (ABN_DISX) captures managers' attempts to boost earnings by reducing discretionary expenditures (DISX), such as advertising, below appropriate levels. In addition, abnormal sales cash flow from operations (ABN_CFO) captures managers' attempts to increase cash flows from operations (CFOs) by accelerating sales through heavier price discounts. The third measure is just the sum of the first two, labeled RM. The measures are computed as follows. [ABN_DISX.sub.i,t] is the predicted discretionary expenses for firm i at time t, minus its actual discretionary expense. Similarly, ABN_CFO,, is the predicted sales cash flow for firm i at time t, minus its actual sales cash flow. Details regarding how the predicted values are estimated appear in Appendix A. The "abnormal" variables are defined in these ways so that a larger value of each implies greater earnings management.

RM accounts for the possibility that when firms reduce discretionary expenses to boost reported income, and impact ABN_DISX, there can be offsetting effects on ABN_CFO (Roychowdhury, 2006). For example, a reduction in advertising can boost reported income, but it may also reduce sales. In order to capture the aggregate effects, we follow Cohen and Zarowin (2010) and Zang (2012) in defining RM as follows:

RM = ABN_DISX + ABN_CFO. (4)

C. Earnings Management and the Cost of Debt

To examine the relation between earnings management and the cost of debt, we model the cost of debt as a function of firm- and issue-specific characteristics (Reeb, Mansi, and Alee, 2001; Anderson, Mansi and Reeb, 2003; Bhojraj and Sengupta, 2003). We then augment that model with the funded status of firms' pension plans and MCs (Campbell et al., 2012) and earnings management. Consistent with prior research, we measure the cost of debt as the difference between yields on the firm's bond issue and a US Treasury issue of similar maturity.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (5)

In the above equation, all firm characteristics are measured at the end of the year preceding the bond issuance. The estimate on MC (funded status) is expected to be positive (negative) if MCs (funded status) are positively (negatively) associated with firms' cost of debt (Campbell et al., 2012). To identify the debt-cost effect of earnings management conditioned on MC shocks, we include an interaction variable between MC and earnings management in some specifications.

D. Earnings Management and the Implied Cost of Equity and WACC

We model the cost of equity as a function of firm characteristics (Botosan and Plumlee, 2005; Dhaliwal, Heitzman, and Li, 2006; Hail and Leuz, 2006; Pastor, Sinha, and Swaminathan 2008). We use the implied cost of equity estimate as some recent research suggests that this measure can better capture the time variation in expected stock returns than ex post realized returns (Pastor et al., 2008). Hail and Leuz (2006) suggest that another advantage of the implied cost of equity model is that it explicitly controls for cash flow and growth effects, which helps to isolate the discount rate effect. We then augment this model with the funded status of firms' pension plans, MCs (Campbell et al., 2012), and earnings management.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (6)

The implied cost of equity is estimated as the internal rate of return that equates the current stock price with the present value of all future cash flows to common shareholders (Gebhardt, Lee, and Swaminathan, 2001). Similar to Pastor et al. (2008) and Campbell et al. (2012), we use the Gebhardt et al.'s (2001) model to estimate the implied cost of equity capital. This model equates the discounted future cash flows to shareholders to the firm's stock price, and then solves for the discount rate (ras) in the following equations:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (7)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (8)

where [P.sub.t] is the price per share of common stock at the end of year t; [B.sub.t] is the book value at the end of year t divided by the number of common shares outstanding at the end of year t; [FEPS.sub.t+i] is the forecasted earnings per share for year t + i; PEPSi and FEPS2 are equal to the one- and two-year-ahead consensus EPS forecasts; FEPS3 is equal to the three-year-ahead consensus EPS forecasts when available, and [FEPS.sub.2] (1 + LTG) when not available. [FROE.sub.t+i] is the forecasted return on equity (ROE) for period t + i. For Years 1 to 3, this variable is equal to [FEPS.sub.t+i]/[B.sub.t+i-1]. Beyond Year 3, [FROE.sub.t+i] is a linear interpolation to the industry median ROE. Industry median ROE is defined as the moving median ROE for the prior 5 to 10 years for the firm's industry (excluding loss firm years). Industries are defined using the 48 industry classifications in Fama and French (1997). Variable [B.sub.t+i] is [B.sub.t+i-1] + [FEPS.sub.t+i] (1-k), T is the forecast horizon, T = 12, and TV is the terminal value.

Following Gebhardt et al. (2001), Dhaliwal et al. (2006), and Campbell et al. (2012), we estimate the cost of equity capital for a given firm year by using analysts' forecasts of the firm's earnings per share in the Institutional Broker's Estimate System (I/B/E/S) data as of June in the following year. As in Campbell et al. (2012), we winsorize the estimated implied cost of equity from above at 0.50. Finally, similar to Gebhardt et al. (2001), we exclude firm-year observations with a negative return on equity from our sample.

As in Equation (5), to identify the equity-cost effect of earnings management conditioned on MC shocks, we include an interaction variable between MC and earnings management in some specifications of Equation (6). We use a model similar to Equation (6) to test for a relation between WACC and MCs.

Following the cost of capital models, we examine the separate relations between MCs and debt issuance, equity issuance, and firm investment. Those models include different control variables from Equations (5) and (6) based on earlier literature that establishes the major drivers of debt issuance, equity issuance, and firm investment. Like Equations (5) and (6), we include variables to control for DB funding status, MCs, earnings management, and, most importantly, an interaction variable to measure the effect of earnings management conditioned on MCs. The sign and significance of the coefficient estimate on the interaction variable tests whether earnings management associated with MCs has a significant impact on debt issuance, equity issuance, and firm investment.

II. Data Description, Empirical Results, and Discussions

A. Data Description

Our sample includes firms in Compustat that sponsor DB plans and have nonmissing data to estimate the earnings management measures. Since MC is determined based on the DB planlevel funding status, we obtain DB plan data from Form 5500s that firms file with the Internal Revenue Service (IRS) annually. Boston College's Center for Retirement Research has compiled these data from 1991 to 2007 and made them available on its website (http://crr.bc.edu/data/ form-5500-annual-reports/). We obtain Form 5500 data from 2008 to 2013 from the US Department of Labor's website (http://www.dol.gov/ebsa/foia/foia-5500.html).

We aggregate MCs, plan assets, liabilities, and actual employer contributions from plan- to firm-level using the CUSIP identifiers from 1991 to 1998 and the Employer Identification Numbers (EINs) for the remaining years when CUSIPs were no longer reported in Form 5500s. Then, we match the firm-level DB pension data with Compustat data based on CUSIPs, EINs, and company names. We acknowledge that our matching process is not perfect, particularly when it is based on EINs and company names, as a firm's subsidiary may have its own EIN number and file a separate Form 5500 or have a different name. In that case, it is possible that the subsidiary's pension data are not aggregated to its parent firm's data, resulting in an underestimate of MCs, pension assets, liabilities, and employer contributions for some firms in our sample. We provide a description of DB plan data aggregation from Form 5500s in Appendix C.

We obtain accounting data from Compustat, stock price and return data from the center for Research in Security Prices (CRSP), cost of debt from the Securities Data Commission (SDC) New Issues database, and analysts' earnings estimates from I/B/E/S over the sample period. We obtain audit firms and tenure data from Audit Analytics (these data are available from 1999 to 2013). Finally, we manually collect the number of audit committee members and audit committee meetings each year for a subsample of bond issuers that file the DEF 14A proxy reports with the Securities Exchange Commission (SEC), which are archived in the Electronics Data Gathering, Analysis, and Retrieval (EDGAR) database. Our final sample spans the period from 1991 to 2013 and includes 1,283 unique firms.

We provide the summary statistics of our sample in Table I. The average firm in our sample is large (with market capitalization of $7.68 billion), profitable (return on assets equals 3.8%), has overfunded DB plans (average funded status equals 0.01), and has an investment-grade credit rating (the average Moody rating score equals 9.17; a rating score of 10 or lower indicates investment grade). Consistent with Rauh (2006), the average ratio of mandatory contributions to firm assets (MC) is 0.002, and is heavily skewed to the right (only firm-year observations in the 75th percentile or above have positive MC values). In addition, the data indicate that our sample firms tend to have lower discretionary accruals relative to their peers as evidenced by the negative values of the performance-adjusted discretionary accruals ratios (means of PDA 1 and PDA2 are -0.007) and manage discretionary expenditures and sales cash flows downward as evidenced by the negative mean values (-0.005 and -0.017, respectively), but manage production costs upward (mean value equal 0.313).

B. Defined Benefits Pension Plan and Earnings Management

Table II reports the NEM and CEM regression results in Panels A and B, respectively. Regressions in Columns 1 and 3 of Panel A covering NEM measures indicate that the coefficient estimates on Funded Status and MC are statistically insignificant. The results in Column 5 indicate that both funded status and MCs are negatively related to ABN_PROD. To address the possibility that collinearity between Funded Status and MC drive our results, we exclude Funded Status from the regression models reported in Columns 2, 4, and 6, but the coefficients on MC are either insignificant or negative suggesting that MCs do not lead firms to engage in noncash-generating earnings management. Indeed, the significant negative estimate on ABN_PROD implies that firms reduce that method of earnings management in response to MCs.

The estimates on the control variables indicate that PDAs are negatively (positively) associated with firm age and financial leverage (firm size and sales growth volatility). In addition, ABN_PROD is positively (negatively) related to firm size proxied by book assets (growth opportunities proxied by the market-to-book ratio).

Panel B of Table II reports the effects of MCs on CEM while controlling for funded status and other variables. The estimates on MC in all three columns are positive and significant at the 1% and 5% levels, respectively, indicating that larger MCs are associated with more aggressive management of discretionary expenditures and sales cash flow.

These results are intuitively appealing. Curbing discretionary expenditures and accelerating sales generate cash that can be used to meet the required pension contributions or make investments. (5) Accelerating production, however, uses cash, even though it spreads fixed costs over more units to boost the profit margin and reported income. Thus, the negative estimate on ABN_PROD is consistent with the notion that firms prefer an earnings management method that generates internal cash, at least in response to MCs. Since ABN_PROD uses cash, firms employ less of that method. The effect of MCs on CEM is economically large. For example, the point estimates in Columns 1 and 2 indicate that an increase of one standard deviation in MCs is associated with an increase of 0.006 and 0.007 in ABN_DISC and ABN_CFO values, respectively. The estimated change is substantial relative to the unconditional means of ABN_DISX and ABN_CFO (-0.005 and -0.017). Given the book asset value of the average sample firm of $7,775 billion, the increase in ABN_DISX and ABN_CFO is equivalent to approximately $46.65 and $54.43 million, respectively.

The negative and significant estimates on Funded Status in Columns 1 and 3 indicate that firms with better funded pension plans use less CEM. To rule out the possibility that collinearity between Funded Status and MC drive our estimation results, in an unreported analysis, we rerun regressions with either Funded Status or MC, but not both in the same model and find qualitatively similar results. In addition, we control for the possible differential effects of the overfunded and underfunded status of DB plans, where funded status is positive or negative, respectively, in our regressions, but our findings persist.

Rauh (2006) reports that the negative real effects of MCs is concentrated in financially constrained firms with low credit ratings. Although MCs represent a shock to internal liquidity that exacerbates financial constraints, there could be firms that are financially unconstrained, but underfund their DB plans (i.e., firms may use underfunding as a tactic during labor contract negotiations with unions) leading to positive MCs. To account for this possibility, we sort firms into subgroups with the S&P's long-term investment or noninvestment-grade credit ratings, and examine the effects of MCs on the earnings management of each subgroup. Our analysis results, reported in Table III, indicate that MCs have significantly stronger (weaker) effects on the CEM of the subgroup with noninvestment (investment) credit ratings. However, MCs have little effect on the NEM of either subgroup (the results for NEM are not reported for brevity). Overall, our evidence indicates that firms that face MCs resort to CEM rather than NEM.

C. Earnings Management and Costs of Capital

Thus far, we have demonstrated the differential effects of MCs on NEM and CEM. In this section, we examine the implications of these effects on DB firms' cost of debt and implied cost of equity. MCs associated with underfunded pension plans drain internal resources that could otherwise be invested in valuable investment projects. Faced with MCs, firms are more likely to need external funds to support pension contributions and corporate investments.

External financing is costlier than internal financing due to information asymmetries, agency costs, incomplete contracting, taxes, and issuance costs (Jensen and Meckling, 1976; Myers, 1977; Myers and Majluf, 1984; Poterba and Summers, 1984). Moreover, the premium required on external financing should be more pronounced for firms that face greater external financing constraints (Almeida, Campello, and Weisbach, 2004; Rauh, 2006; Franzoni, 2009; Campbell et al., 2012). Campbell et al. (2012) empirically examine the impacts of MCs (funded status) on the costs of capital and report that MCs (funded status) increase (reduce) the cost of debt, the implied cost of equity, and the WACC. We now examine whether earnings management associated with MCs mitigates or exacerbates the increased costs of capital caused by MCs.

Earlier research regarding the effects of earnings management on the cost of capital is mixed. Trueman and Titman (1988), Chaney and Lewis (1995), Guay, Kothari, and Watts (1996), Subramanyam (1996), Demski (1998), and Arya, Glover and Sunder (2003) suggest that earnings management can benefit firms as it can reduce their perceived probability of bankruptcy or signal other positive information to investors, thereby lowering borrowing rates. Conversely, Leuz and Verrecchia (2005) argue that poor quality reporting impairs the coordination between firms and their investors with respect to the firm's capital investment decisions. This impairment increases the cost of equity (Francis et al., 2004; Aboody, Hughes, and Liu, 2005) as well as the cost of debt (Francis et al., 2005; Bharath, Sunder, and Sunder, 2008). However, none of these studies identifies the effects of earnings management on the costs of capital using MCs as an exogenous shock. We explore this area next.

Table IV reports the regression results for the effects of earnings management, MCs, and their interaction on the cost of debt. The estimates on the interaction between earnings management measures and MCs represent the effect of earnings management, conditional upon an MC shock. We examine the effects of NEM in Panel A and CEM in Panel B. Firm characteristic variables are measured at the beginning of the year. First, we replicate the cost of debt regression of Campbell et al. (2012) in Column 1 of Panel A. Like Campbell et al. (2012), we find that MC is positively related to the cost of debt (estimate = 0.16, p-value = 0.10). (6) The estimates on PDA1, PDA2, and ABN_PROD are negative, but insignificant indicating that the effect of NEM on the cost of debt is not statistically different from zero. When we include the interaction between the measures of NEM and MC in Columns 3, 5, and 7, the estimates on the interactions between PDAs and MC are positive and statistically significant implying that creditors raise their return premiums for MC firms that use noncash generating earnings management. The economic impact on the cost of debt of MC firms is relatively small. Using the point estimates in Column 3 for illustration, with the control variables set at their sample means, the total effect of a one standard deviation increase in PDA 1 on the cost of debt for a firm with the average MC is only minus one basis point (= -0.006 x 0.131 + 2.471 x 0.131 x 0.002).

Panel B reports some similar results for the effects of CEM. Although the estimates on discretionary expenditures (ABN_DISX) management as reported in Columns 1 and 2 are insignificant, the estimate on the interaction between MC and ABN_CFO in Column 4 is positive and statistically significant. This result suggests that creditors view the firms that incur MCs and aggressively manage CFOs as particularly risky and demand a premium yield. Similarly, the estimate on the interaction between MC and RM is positive and significant in Column 6. The point estimates in Column (4) indicate that a one standard deviation increase in ABN_CFO results in a decrease of only two basis points (i.e., 0.02%) in the cost of debt of an MC firm. The decrease occurs because the average negative effect of ABN_CFO swamps the marginal positive effect of ABN_CFO x MC computed at the variable means.

Collectively, our evidence indicates that earnings management increases the cost of debt, but only when conditioned on MCs. That is, when creditors know that a firm must make large MCs and that firm also manages earnings, they demand a higher return on the firm's debt. The firm's creditors could view MC-driven earnings management more negatively as MC funding draws away cash that would otherwise be available to support their claims. Furthermore, given the negative long-term effects associated with earnings management, creditors may interpret earnings management in the presence of MCs as a signal of an MC firm's financial distress. (7)

Table V reports the effects of NEM and CEM on the implied cost of equity in Panels A and B, respectively. Column 1 of Panel A replicates the implied cost of equity regression in Campbell et al. (2012). Columns 2 to 7 test the effects of NEM on the implied costs of equity. The estimates on MC across the columns are positive and statistically significant, which is similar to those reported by Campbell et al. (2012). Furthermore, we find that NEM has a significant and positive average impact on the implied cost of equity and the various methods of NEM do not help to reduce the equity cost of MC firms.

The results in Panel B indicate that CEM has a positive, but usually insignificant average effect on the implied cost of equity. For example, a one standard deviation increase in ABN_DISX increases the cost of equity by approximately 0.02%. CEM associated with MCs, however, may reduce the implied cost of equity a bit, although only the estimate on the interaction MC x ABN_DISX is negative and statistically significant. Holding other control variables unchanged at their sample means, a one standard deviation increase in ABN_DISX results in an increase of one basis point (0.01%) of the implied cost of equity of the MC firms, which is trivial.

Our evidence thus far indicates that MC firms that manage earnings more aggressively generally incur higher costs of debt, but cannot lower their costs of equity much. Next, we examine the relations between WACC, earnings management, and MCs, while controlling for other variables.

We model WACC as a function of firm- and issue-specific characteristics, along with our earnings management and MC variables, and report the results in Table VI. We use the estimates of marginal tax rates developed by Blouin, Core, and Guay (2010) in calculating WACC. Most of the estimates regarding the earnings management and interaction variables are statistically insignificant except for those on the interactions between MC and either ABN_DISX or RM. The significant negative estimates imply that discretionary expenditures management conditioned on MCs reduces WACC. Nevertheless, the positive average effect of ABN_DISX on WACC is large enough to make the net effect of discretionary expenditures management on the WACC of MC firms very small.

In summary, our investigation on the effects of earnings management on the cost of capital indicates that CEM through discretionary expenditures conditioned on MCs can offset some of the positive average effect of MCs on firms' implied cost of equity and WACC, but the economic effect is small.

D. Earnings Management and External Financing

The results above confirm that earnings management does not help firms to reduce the cost of capital significantly implying that it provides little help when seeking to raise external capital to cover pension shortfalls and fund investments. In this subsection, we examine the direct effects of earnings management on debt and equity financing and report the results in Panels A and B, respectively, of Table VII.

Net debt issuance is measured as the difference between long-term debt issuance and long-term debt reduction, all scaled by book assets at the beginning of the year. Net equity issuance is constructed as the difference between the sale of common stock and stock repurchased, all scaled by book assets at the beginning of the year. We follow Campello and Graham (2013) and Linck et al. (2013) in selecting control variables for the external financing regression models. Both the test and control variables are measured at the end of the preceding year.

The results in Panel A of Table VII demonstrate that the estimates on earnings management measures and their interactions with MC are statistically insignificant indicating that earnings management has little effect on new debt issuance. The results in Panel B report a negative relation between PDAs and equity issuance conditioned on MCs. This result is consistent with our finding of a positive relation between PDAs and the implied cost of equity. The remaining types of earnings management have insignificant effects on external equity financing. These results are consistent with our cost of capital analysis. If earnings management has little impact on a firm's cost of capital, then it is unlikely to help the firm raise more external financing, all else being equal. (8)

E. Earnings Management and Real Investments

Linck et al. (2013) argue that earnings management signals positive prospects about firm profitability, which helps to raise external capital to finance their investments. Cohen and Zarowin (2010) find that firms manipulate earnings prior to seasoned equity offerings (SEOs). In this section, we consider the effects of earnings management on firm investment using MCs to identify the relation.

If earnings management mitigates information asymmetry between managers and investors by signaling profitability, it could facilitate external capital financing and help increase investment.

This would be particularly useful for those firms with positive MCs as these firms must divert their limited internal resources from profitable investment projects to pension contributions (Rauh, 2006). However, if earnings management simply adds noise or signals financial distress, it could do the opposite and decrease investment.

We use the conventional investment model that relates investment to internal cash flows while controlling for investment opportunities proxied by Tobin's Q (Fazzari, Hubbard, and Peterson, 1988). Note that the conventional investment model could be subject to measurement errors as Tobin's Q captures the average investment opportunity, not the desired marginal investment opportunity. The unobserved marginal investment opportunity could correlate with the cash flow variable biasing the model coefficient estimates. Therefore, we follow Rauh (2006) in augmenting the model with funded status and MCs, and further include earnings management and an interaction between earnings management and MCs as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (9)

Investment is measured as capital expenditures scaled by book assets at the beginning of the year. Nonpension cash flow is calculated as the sum of net income, depreciation, and amortization expense and pension expenses, all scaled by book assets at the beginning of the year.

Rauh (2006) suggests that corporate investments should not be associated with MCs after controlling for Tobin's Q, cash flow, and funded status except when MCs capture a direct response of investment to exogenous shocks in internal resources. He reasons that the sharp nonlinearities of pension funding requirements, particularly around the threshold of underfunding, allow for the identification of an effect of MCs on investment that is purged of the potentially endogenous relation between funded status and the firm's unobserved investment opportunities.

Table VIII, Panel A, reports the investment regression with NEM and CEM and firm-level fixed effects. The signs and significance of the coefficient estimates of MC and Funded Status across the specifications are consistent with those documented by Rauh (2006) and Campbell et al. (2012). Earnings management variables are mostly statistically insignificant except ABN_CFO, which is positive (0.001) and significant at the 1% level. However, the coefficient estimate of the interaction ABN_CFO x MC is negative (-0.335) and highly significant suggesting that CFO management exacerbates the negative impact of MCs on corporate investment.

It is possible that firms make decisions on earnings management and corporate investment jointly raising an endogeneity concern that could bias our coefficient estimates. To alleviate this endogeneity concern, we use the nonlinear generalized method of moments (GMM) to estimate Equation (10) and report the results in Panel B of Table VIII. We use nonlinear estimation as our model includes the product of an endogenous variable and an exogenous variable. Thus, a linear instrumental variable (IV) estimator would produce inconsistent estimates (Billet, King and Mauer, 2007). We follow Zang (2012) in selecting instruments for earnings management. Specifically, we use auditor quality, auditor tenure, post-SOX dummy, and operating cycle variables as instruments for discretionary earnings management. Auditor quality is a dummy variable that takes a value of one if the audit firm is one of the Big 8 (or Big 6, Big 5, or Big 4 in recent years) audit firms, and zero otherwise. Auditor tenure is the number of years that the audit firm has continuously worked as an auditor for a client firm. (9) To instrument for CEM, we use market share and marginal tax rates. Zang (2012) argues that firms with larger market shares have greater flexibility in deviating from the optimal operating decisions due to their competitive advantage in the industry. Alternatively, the tax costs associated with CEM are higher for those firms with higher marginal tax rates implying a negative relation between the marginal tax rates and CEM. Overall, the selected instruments are likely to explain the respective measures of earnings management, but unlikely to have direct effects on corporate investment.

Before performing the GMM estimation, we conduct endogeneity tests of the earnings management variables, tests of weak instruments, and tests of instrument validity (these test results are not reported, but are available on request). The Hausman endogeneity tests validate our endogeneity concern for PDA1, PDA2, ABNJDISX, and RM. For theses variables, the ordinary least square (OLS) estimates are biased and inconsistent. The weak instruments and instrument validity tests suggest that our selected instruments are appropriate (the Cragg-Donald Wald F-statistics > 10 and the Hansen J-statistics are insignificant). (10)

The GMM estimation results indicate that among the earnings management variables, PDA1, PDA2, ABN_PROD, and ABN_DISX are negatively and significantly related to corporate investment. However, none of the estimates on the interactions of MC and the measures of earnings management is statistically significant with the exception of the one between ABN_CFO and MC, but it is negative. These results indicate that earnings management does not mitigate the negative effects of MCs on investment. (11)

Overall, our analysis finds little support for the argument that earnings management mitigates the negative real effects of MCs. The results are consistent with our earlier analysis of the cost of capital and external financing's effects on earnings management. If earnings management does not help financially constrained firms raise significant additional external financing, it is unlikely to help firms increase investment. In addition, the result could reflect the cash use priority that DB firms place on pension contribution obligations. Finally, it could also be consistent with earlier research that reports that financially constrained firms build up liquidity and recapitalize in the short-term before embarking on new investment programs in the long-term (Pulvino and Tarhan, 2006; Almeida and Campello, 2010; Marchica and Mura, 2010; Dasgupta, Noe, and Wang, 2011; Phan and Hegde, 2013).

F. Earnings Management and Pension Contributions

If earnings management has little effect on the financing and investments of MC firms and is costly to them, one may ask why MC firms still manage earnings, particularly CEM. To answer this question, we examine the relation between earnings management and actual pension contributions of DB plan firms.

Because our earlier analysis indicates a positive relation between MC and CEM, we focus our analysis in this section on a subsample of firms that have positive MCs in the previous year. Specifically, we use an independent double sort to first sort DB firms into Large MC and Small MC subgroups based on their previous year's ratio of MCs to the book value of assets relative to the subsample median ratio. Then, we sort these firms into the High EM and Low EM subgroups based on their NEM and CEM measures, relative to the respective sample medians.

We report the univariate analysis results of the difference in the means of DB plan contributions of the High EM and Low EM subgroups, both belonging to the Large MC subgroup, in Panel A of Table IX. Using the t-test for statistical inference (the unreported Wilcoxon rank-sum tests provide qualitatively similar results), we find that between two firms that incur large MCs, the one that uses CEM more aggressively is able to make larger contributions to its DB plan(s). This is true for all three measures of CEM. However, there is no statistically significant difference when we use any of the three NEM measures to capture the effects on plan contributions.

This evidence is consistent with our earlier finding of a positive relation between MCs and CEM. However, our analysis of the Small MC subgroup reported in Panel B of Table IX reflects little difference in DB contributions. In sum, our evidence suggests that CEM helps MC burdened firms contribute more to their underfunded DB plans.

III. Robustness Checks

We conduct a battery of robustness checks of our findings. For brevity, the results are not reported, but are available from the authors. Zang (2012) finds that managers use the various earnings management methods as substitutes for one another. As such, we have tested the effect of these forms of earnings management separately. Since firms may use several forms of earnings management concurrently, we include all of the methods, as well as their interactions with MCs in our cost of capital, external financing, and investment regressions, but our findings are qualitatively unchanged. In addition, we estimate the effects of MCs on NEM and CEM in a system of simultaneous equations, but our results still hold.

The acceleration of the amortization of funding shortfalls from 30 to seven years beginning in 2008 as stipulated by the Pension Protection Act (PPA) 2006, coupled with the 2008 to 2009 recession, could amplify the impact of MCs and exacerbate financial constraints. To account for a possible structural break in our estimation, we construct a dummy variable that takes a value of one for any year after 2007, and zero otherwise. We augment previous regression models with interactions of this dummy variable and each of the test variables and rerun our analyses. We find that the effect of MCs on earnings management or the interaction effect of MCs and earnings management on the outcome variables between the pre- and post-2008 periods does not change significantly.

SOX in 2002 could have changed firm preferences for NEM versus CEM. To examine the possible effect of SOX on the earnings management of MC firms, we split the sample into pre- and post-SOX subsamples and rerun our tests for each subperiod. We do not find a significant effect of MCs on NEM in either the pre- or post-SOX periods. However, we find that MCs have a positive effect on CEM in the pre-SOX period, but the relation weakens in the post-SOX period. The effects of CEM on MC firms' financing and investments do not change. This evidence further corroborates our earlier finding that CEM, which can generate immediate cash, is more important to firms burdened with MC obligations.

Although the cost of capital and the amounts of external financing provided to firms are likely to be determined by the markets, one cannot completely rule out the possibility that firms manage earnings in order to obtain some target levels of external financing or cost of capital or both. This raises the possibility of reverse causality and endogeneity. To address this concern, we re-estimate the cost of capital and external financing models using GMM and employing instruments for earnings management similar to those used earlier, but the GMM estimation results are qualitatively similar to the OLS results.

Campbell et al. (2012) find that the positive effect of MCs on the cost of capital is more pronounced for firms with poor credit ratings. Therefore, in an additional analysis, we examine the effects of MCs on the cost of capital, financing, and investment of each subgroup. Our results indicate that MCs have more pronounced effects on the cost of capital of firms with noninvestment ratings, but earnings management does not help them with financing or investment. However, we find that sales cash flow management is positively related to actual pension contributions for both subgroups. In an alternative analysis, we follow Hadlock and Pierce (2010) and sort firms into financial constraint subgroups based on a size-age index and rerun the test for each subgroup, but the results are qualitatively unchanged.

Note that pension liabilities account for 17.73% (10.12%) of an average (median) firm's book value of assets and only a portion (approximately 25%) of the sample firms incur positive MCs in a given year. In addition, the matching of Form 5500 data and those in Compustat is not perfect, resulting in differences between firm-level pension assets and liabilities aggregated from Form 5500 and those reported in Compustat. Therefore, to focus our analysis on the DB pension plans that are important to the sponsoring firms' operations and to mitigate the data discrepancy issue, we drop those firm-year observations where the ratio of aggregated pension liabilities to book assets is in the bottom half of the sample. We additionally drop those firm-year observations where the ratio of aggregated pension liabilities to the pension obligations reported in Compustat is below 50% or above 150%. The means of MC and the ratio of aggregated pension liabilities to the book assets of firms in the new sample increase by 61% and 70% over the full sample's means, respectively. However, we find that our results based on this new sample are qualitatively unchanged.

Following the adoption of the SFAS 132(R) accounting standard in 2003, DB plan sponsors are required to report their expected pension contributions in the upcoming fiscal year on the annual 10-K report (Compustat variable #PBECE). This measure is first used by Kubick, Lockhard, and Robinson (2014) in their research on the effect of the change in internal liquidity on DB plan sponsoring firms' stock value and investment following the pension funding relief provided by the Moving Ahead for Progress in the 21st Century Act (MAP21). Because this measure is reported by firms, it is probably less prone to the underestimation of pension assets and liabilities associated with our aggregation of pension data from plan- to firm-level. However, this measure poses two potential challenges to our analysis: 1) it is only available for fiscal years ending after December 15, 2003, resulting in a smaller sample size and 2) the expected pension contributions may include both mandatory and discretionary pension contributions and it is hard to separate the two. Nevertheless, we use this measure as a proxy for MCs in our next robustness check and note that our findings are qualitatively unchanged. (12)

Bakke and Whited (2012) suggest that the regression discontinuity design (RDD) can only alleviate endogeneity problems for those observations close to the threshold. These authors reexamine Rauh's (2006) investment analysis and report that the negative effect of the exogenous shock to internal cash flow induced by MCs on capital expenditures documented by Rauh is concentrated in a small group of heavily underfunded firms that differ significantly from the rest of the firms. Alternatively, they find that MCs have causal effects on R&D investment, account receivables, and hiring. In a robustness check, similar to Bakke and Whited (2012), we substitute a firm's funded status and MCs with distance from an underfunded point (e.g., 0%, 10%, or 20%) and an indicator variable that takes a value of one if the firm's funding status is below the selected underfunded point, and zero otherwise and rerun our analyses. Our results indicate that the positive and significant relation between underfunding status and CEM persists given different underfunded points. In addition, our other findings are qualitatively unchanged. (13)

IV. Conclusions

This paper examines whether MC driven shocks to firms' internal liquidity induces them to manage earnings more aggressively, either through NEM or CEM. We test whether earnings management induced by MCs has beneficial effects on firms' cost of capital, external financing, and corporate investment. We find that firms respond to MCs by increasing CEM, particularly by managing discretionary expenditures and sales cash flows more aggressively, but do not increase NEM. We determine that, on average, MCs increase the cost of debt and any earnings management associated with MCs only exacerbates that increase. MCs also increase the cost of equity, on average, and earnings management does little to counterbalance that increase. The same holds true for the WACC. Moreover, earnings management induced by MCs has little effect on corporate financing and investments. Finally, we find evidence that CEM generates cash that helps DB sponsoring firms to meet their pension contribution obligations.

Appendix A: Details of Earnings Management Variable Construction

Variables Used to Compute Accrual-Based Earnings Management Measures

We follow Kothari, Leone, and Wasley (2005) in calculating total accruals (7X) as follows:

[TA.sub.i,t] = ([DELTA][CA.sub.i,t] - [DELTA][CL.sub.i,t] - [DELTA][Cash.sub.i,t] + [DELTA][STD.sub.i,t] - [Dep.sub.i,t])/[Asset.sub.i,t-1], (1A)

where [TA.sub.i,t] is the total accruals of firm i at time t, and [DELTA] represents a one-year change in a variable. CA is current assets, CL is current liabilities, Cash is cash holdings, and STD is long-term debt in current liabilities, Dep is the depreciation and amortization expense of the firm.

A portion of total accruals is "nondiscretionary" and beyond the control of the managers. Thus, we follow Dechow, Slow, and Sweeney (1995) in separating nondiscretionary from discretionary accruals (DAs). Specifically, we use a version of the Jones (1991) model of accruals to estimate nondiscretionary accruals (NDAs) as the predicted value from a regression of total accruals on the inverse of lagged firm size, the change in firm sales, and gross property, plant, and equipment scaled by total firm assets as follows:

[TA.sub.i,t] = [[alpha].sub.0] + [[alpha].sub.1] (1/[A.sub.i,t-1]) + [[alpha].sub.2] ([DELTA][REV.sub.i,t]) + [[alpha].sub.3] ([PPE.sub.i,t]) + [[epsilon].sub.i,t]. (2A)

The variable [DELTA][REV.sub.i,t] is the change in sales, normalized by lagged assets, for firm i at time t and [PPE.sub.i,t] is gross property, plant, and equipment, also normalized by firm assets.

Similar to Cohen and Zarowin (2010), we run cross-sectional regressions for each of the Fama-French (1997) 48 industries annually to estimate [[alpha].sub.0], [[alpha].sub.1], [[alpha].sub.2], [[alpha].sub.3] in Equation (2A) and use the estimates and firm data to predict the level of NDAs for each firm-year observation. The predicted values, [[??].sub.i,t], are used in Equation (1) in the text to compute a firm's discretionary accruals.

Our second measure of nondiscretionary accruals (NDA2), used to compute discretionary accruals DA2, uses a version of the "modified Jones" model. It replaces the change in sales in Equation (2A) with the change in sales less the change in receivables (REC) to estimate NDA2 as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (3A)

Variables Used to Compute Real Earnings Management Measures

Following prior research (Roychowdhury, 2006; Cohen and Zarowin, 2010), we estimate the normal levels of discretionary expenditures, cash flows from operation, and production costs as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (4A)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (5A)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (6A)

DISX represents the discretionary expenditures, defined as the sum of SG&A, advertising expenses, and R&D expenses, CFO is the sales cash flow from operations, and PROD is the production costs, which are defined as the sum of the costs of goods sold (COGS) and the change in inventories. We estimate Equations (4A) to (6A) for each of the Fama-French (1997) 48 industries annually. Each sample firm is assigned to one of the 48 industries, and the estimates from these regressions are used to compute the predicted values of DISX, CFO, and PROD for each firm for each time t.
Appendix B: Variables Construction

Variable Name              Variable Construction       Data Source

Dependent variables

Yield Spread           The yield on the first bond     SDC Platinum
                       issue in year t less the
                       yield on a corresponding
                       Treasury security of similar
                       maturity divided by 1,000.

Net Debt Issuance      Net debt issuance is            Compustat
                       measured as the difference
                       between the long-term debt
                       issuance and the long-term
                       debt reduction, all divided
                       by the total assets at the
                       beginning of the year.

Net Equity Issuance    Net equity issuance is          Compustat
                       measured as the difference
                       between the sale of common
                       stock and stock repurchases,
                       all divided by the total
                       assets at the beginning of
                       the year.

Investment             Capital expenditures divided    Compustat
                       by the book assets at the
                       beginning of the year.

Independent Variables

Funded Status          Pension assets minus pension    Form 5500s
                       liabilities, both at the
                       plan-level, aggregated to
                       the firm-level and divided
                       by the market value of
                       equity as of year t-1.

Mandatory              Max (MFC, DRC), where MFC is    Form 5500s
Contribution           the minimum funding
                       contribution and DRC is the
                       deficit reduction
                       contribution. MFC is
                       approximated as the normal
                       cost due during the year
                       plus 10% of the previous
                       period's funding gap. DRC as
                       a fraction of the funding
                       gap is calculated as min
                       (0.30, [0.30-0.25 (pension
                       assets/pension liabilities
                       --0.35)]) of the
                       underfunding amount from
                       1991 to 1994, and as min
                       (0.30, [0.30-0.40 (pension
                       assets/pension liabilities
                       --0.60)]) of the
                       underfunding amount for 1995
                       and later. Mandatory
                       contributions are calculated
                       at the plan-level then
                       aggregated to firm-level.

Market-to-Book         Market-to-book ratio is the     Compustat
Ratio                  ratio of the market value of
                       assets to the book value of
                       assets.

Log(Market Value of    The natural log of the          Compustat
Equity)                market value of equity
                       reported by CRSP at the end
                       of year t-1.

Book Leverage          Short-term debt plus            Compustat
                       long-term debt divided by
                       total assets.

Debt-to-Equity         Long-term debt at the end of    Compustat
Ratio                  year t-1 plus the proceeds
                       of the first debt issue
                       after the end of year t-1,
                       divided by the market value
                       of equity at the end of year
                       t-1.

Book-to-Market         Book value of equity at the     Compustat
Ratio                  end of year t-1 divided by
                       the market value of equity,
                       winsorized at zero and
                       three.

Nonpension Cash        Nonpension cash flow is         Compustat
Flow                   measured as (net income +
                       depreciation and
                       amortization + pension
                       expense) / book assets at
                       the beginning of the year.

Tobin's Q              The sum of the market value     Compustat
                       of equity plus book assets
                       minus book equity minus
                       deferred taxes, all scaled
                       by book assets.

ROA                    Income before extraordinary     Compustat
                       items divided by total
                       assets at the end of year
                       t-1.

Firm Age               The number of years that        Compustat
                       firms have been included in
                       Compustat.

Operating Cycle        Days receivable plus the        Compustat
                       days inventory less the days
                       payable at the beginning of
                       the year.

Credit Rating Score    Moody's Credit Rating in        SDC Platinum
                       year t converted to a
                       numerical equivalent, where
                       one is assigned to bonds
                       with an Aaa rating and 19 is
                       assigned to bonds with a C
                       rating (as reported in the
                       SDC New Issues database).

Analyst Forecast       Analyst forecast dispersion     I/B/E/S
Dispersion             is the natural logarithm of
                       the standard deviation of
                       analyst estimates for the
                       next period's earnings
                       divided by the consensus
                       forecast for next period's
                       earnings

Long-Term Growth       Long-term growth rate is the    I/B/E/S
Rate                   analysts' long-term growth
                       forecast available from the
                       reported I/B/E/S database.

Market Beta            Market beta is the capital      CRSP
                       market beta estimated with
                       the market model using
                       value-weighted CRSP returns
                       and a minimum of 24 monthly
                       returns over the prior 60
                       months.

Industry Cost of       Industry cost of equity is      Compustat
Equity                 the average cost of equity
                       in each year for each
                       industry based on the
                       Fama-French (1997) 48
                       industry classifications.

Stock Return           Stock return volatility is      CRPS
Volatility             the standard deviation of
                       monthly stock returns for
                       the 24 months through the
                       end of the last month of
                       year t-1.

Log (Bond Maturity)    The natural log of the          SDC Platinum
                       number of years until the
                       maturity of the issue.

Log (Debt Proceeds)    The natural log of the total    SDC Platinum
                       amount of proceeds received
                       from the issue.

Public Debt Dummy      Public debt indicator is an     SDC Platinum
                       indicator variable equal to
                       one if the debt is publicly
                       traded debt and zero if
                       issued under Rule 144a.

Senior Debt Dummy      An indicator variable equal     SDC Platinum
                       to one if the debt issue is
                       senior and zero otherwise.

Auditor Quality        A dummy variable that takes     AuditAnalytics
                       a value of one if the audit
                       firm is a Big 8 (or Big 6,
                       Big 5, and Big 4 in recent
                       years) audit firm, and zero
                       otherwise.

Auditor Tenure         Number of years that the        AuditAnalytics
                       auditing company has
                       continuously worked as
                       auditor for a client firm.


Appendix C: DB Pension Plan Data Aggregation from Form 5500s

We use the DB plan data reported in Form 5500s that firms file with the IRS annually to estimate mandatory pension contributions (MCs) and funding status. We aggregate DB plan-level MCs, funding status, pension assets and liabilities, and employer pension contributions data to the firm level using CUSIPs from 1991 to 1998 and EINs from 1999 to 2013 (when CUSIPs were no longer reported on Form 5500s). Our firm-level aggregate data include 5,124 unique CUSIPs with 24,422 CUSIP-year observations from 1991 to 1998 and 99,569 unique EINs with 567,610 EIN-year observations from 1999 to 2013.

For comparison purposes, we merge the CUSIP- and EIN-level DB pension data from Form 5500 with the pension data reported in Compustat based on CUSIP and EIN identifiers from 1991 to 1998 and 1999 to 2013, respectively. We note that a firm's subsidiary may have its own EIN number and file a separate Form 5500. In addition, a firm without subsidiaries may have several plants in different locations and each location can have its own EIN. It is possible that these subsidiaries' or plants' pension data are not aggregated into its parent firm's data, resulting in an underestimate of MCs, pension assets, liabilities, and employer contributions for some firms in our sample. (14) Therefore, to maximize the match rate, we also use company names for matching. Due to the large sample size, we first use the fuzzy matching function in the SAS statistical package to match company names in Form 5500s with those in the Compustat pension data and then verify the match manually. We are able to match 19,300 firm-year observations in Form 5500s with those in the Compustat pension data set from 1991 to 2013. Given the total number of 42,260 firm-year observations with nonmissing pension data in Compustat in the sample period, the success match rate is approximately 46% (the breakdown of the matching success rate is about 36% and 52% for the subperiods 1991 to 1998 and 1999 to 2013, respectively).

To prepare the data for the regression analysis, we match the pension data with other Compustat data and drop those firm-year observations with missing data for accrual earnings management. Imposing this condition leaves us with 1,283 unique firms with 12,773 firm-year observations with pension data from Form 5500s and 2,437 unique firms with 24,987 firm-year observations with Compustat pension data, representing a match rate of 52%. In addition, on average, the firm-year pension liability (asset) values reported in Form 5500s are equivalent to approximately 87% (103%) of those reported in Compustat over the sample period. The following table summarizes the annual number of observations and the annual means of firm-level pension asset and liability data aggregated from Form 5500s and those from Compustat.
             Form 5500 Pension Data

                   Avg.         Avg.
                  pension      pension
                  assets     liabilities
Year       N       ($M)         ($M)

1991      115      380.64       162.45
1992      581      544.65       431.22
1993      608      490.36       380.46
1994      528      597.09       472.25
1995      610      499.03       433.08
1996      588      741.59       602.95
1997      634      766.84       960.85
1998      523      873.57       658.04
1999      599      853.85       701.36
2000      684     1349.47      1030.82
2001      656     1172.71       990.94
2002      640     1175.55      1125.13
2003      635     1159.14      1339.50
2004      616     1228.73      1206.32
2005      614     1329.44      1250.06
2006      603     1550.02      1474.49
2007      555     2057.42      1860.09
2008      538     1631.89      1512.22
2009      521     1301.13      1213.23
2010      507     1443.93      1447.12
2011      506     1395.33      1482.57
2012      500     1636.69      1510.93
2013      412     1873.36      1700.18

              Compustat Pension Data

                    Avg.          Avg.
                   pension       pension
                   assets      liabilities     Match
Year       N        ($M)          ($M)          rate

1991     1,188      554.92        511.31        9.68%
1992     1,222      557.28        533.95       47.55%
1993     1,246      598.64        605.69       48.80%
1994     1,287      563.57        555.13       41.03%
1995     1,277      670.46        658.25       47.77%
1996     1,296      729.67        667.71       45.37%
1997     1,282      820.31        721.93       49.45%
1998     1,222      902.30        817.32       42.80%
1999     1,167     1060.15        845.47       51.33%
2000     1,081     1116.95        928.72       63.27%
2001     1,027     1037.82       1051.25       63.88%
2002     1,010      931.85       1166.08       63.37%
2003     1,005     1102.34       1299.37       63.18%
2004     1,008     1234.73       1431.41       61.11%
2005     1,023     1285.48       1466.00       60.02%
2006     1,042     1414.91       1489.32       57.87%
2007       998     1518.41       1502.76       55.61%
2008       974     1160.45       1492.39       55.24%
2009       941     1204.62       1516.05       55.37%
2010       929     1483.12       1822.95       54.57%
2011       924     1542.92       1996.17       54.76%
2012       926     1621.95       2153.14       54.00%
2013       912     1758.05       2051.79       45.18%


References

Aboody, D., J. Hughes, and J. Liu, 2005, "Earnings Quality, Insider Trading, and Cost of Capital," Journal of Accounting Research 43, 651-673.

Almeida, H. and M. Campello, 2010, "Financing Frictions and the Substitution Between Internal and External Funds," Journal of Financial and Quantitative Analysis 45, 589-622.

Almeida, H., M. Campello, and M. Weisbach, 2004, "The Cash Flow Sensitivity of Cash," Journal of Finance 59, 1777-1804.

Anderson, R., S. Mansi, and D. Reeb, 2003, "Founding Family Ownership and the Agency Cost of Debt," Journal of Financial Economics 68, 263-285.

Arya, A., J. Glover, and S. Sunder, 2003, "Are Unmanaged Earnings Always Better for Shareholders?" Accounting Horizons 17, 111-116.

Bakke, T. and T. Whited, 2012, "Threshold Events and Identification: A Study of Cash Shortfalls," Journal of Finance 67, 1083-1 111.

Bharath, S., J. Sunder, and S. Sunder, 2008, "Accounting Quality and Debt Contracting," Accounting Review 83, 1-28.

Bhojraj, S. and P. Sengupta, 2003, "Effect of Corporate Governance on Bond Ratings and Yields: The Role of Institutional Investors and Outside Directors," Journal of Business 76, 455-475.

Billet, M., T.-H. King, and D. Mauer, 2007, "Growth Opportunities and the Choice of Leverage, Debt Maturity, and Covenants," Journal of Finance 62, 697-730.

Blouin, J., J. Core, and W. Guay, 2010, "Have the Tax Benefits of Debt Been Overestimated?" Journal of Financial Economics 98, 195-213.

Botosan, C. and M. Plumlee, 2005, "Assessing Alternative Proxies for the Expected Risk Premium," Accounting Review 80, 21-53.

Campbell, J., D. Dhaliwal, and W. Schwartz, Jr., 2012, "Financing Constraints and the Cost of Capital: Evidence from the Funding of Corporate Pension Plans," Review of Financial Studies 25, 868-912.

Campello, M. and J. Graham, 2013, "Do Stock Prices Influence Corporate Decisions? Evidence from the Technology Bubble," Journal of Financial Economics 107, 89-110.

Chaney, P. and C. Lewis, 1995, "Earnings Management and Firm Valuation Under Asymmetric Information," Journal of Corporate Finance 1, 319-345.

Cohen, D., A. Dey, and T. Lys, 2008, "Real and Accrual-Based Earnings Management in the Pre- and Post-Sarbanes Oxley Periods," Accounting Review 83, 757-787.

Cohen, D. and P. Zarowin, 2010, "Accrual-based and Real Earnings Management Activities Around Seasoned Equity Offerings," Journal of Accounting and Economics 50, 2-19.

Cornett, M., J. McNutt, and H. Tehranian, 2009, "Corporate Governance and Earnings Management at Large US Holding Bank Companies," Journal of Corporate Finance 15, 412-430.

Dasgupta, S., T.N. Noe, and Z. Wang, 2011, "Where Did All the Dollars Go? The Effect of Cash Flows on Capital and Asset Structure," Journal of Financial and Quantitative Analysis 46, 1259-1294.

Dechow, P.M., R.G. Sloan, and A.P. Sweeney, 1995, "Detecting Earnings Management," Accounting Review 70, 193-225.

Demski, J., 1998, "Performance Measure Manipulation," Contemporary Accounting Research 15, 261-285.

Dhaliwal, D., S. Heitzman, and O. Li, 2006, "Taxes, Leverage, and the Cost of Equity Capital," Journal of Accounting Research 44, 691-723.

DuCharme, L., P. Malatesta, and S. Sefcik, 2004, "Earnings Management, Stock Issues, and Shareholder Lawsuits," Journal of Financial Economics 71, 23-49.

Duchin, R., O. Ozbas, and B. Sensoy, 2010, "Costly External Finance, Corporate Investment, and the Subprime Mortgage Credit Crisis," Journal of Financial Economics 97, 418-435.

Fama, E. and K. French, 1997, "Industry Costs of Equity," Journal of Accounting and Economics 43, 153-193.

Fazzari, S., G. Hubbard, and B. Petersen, 1988, "Financing Constraints and Corporate Investment," Brookings Papers on Economic Activity 1, 141-195.

Francis, J., R. LaFond, P. Olsson, and K. Schipper, 2004, "Cost of Equity and Earnings Attributes," Accounting Review 79, 967-1010.

Francis, J., R. LaFond, P. Olsson, and K. Schipper, 2005, "The Market Pricing of Accruals Quality," Journal of Accounting and Economics 39, 295-327.

Franzoni, F., 2009, "Underinvestment vs. Overinvestment: Evidence from Price Reactions to Pension Contributions," Journal of Financial Economics 92, 491-518.

Gebhardt, W., C. Lee, and B. Swaminathan, 2001, "Toward an Implied Cost of Equity," Journal of Accounting Research 39, 135-176.

Graham, J.R., C.R. Harvey, and S. Rajgopal, 2005, "The Economic Implications of Corporate Financial Reporting," Journal of Accounting and Economics 40, 3-73.

Guay, W., S. Kothari, and R. Watts, 1996, "A Market-Based Evaluation of Discretionary Accruals Models," Journal of Accounting Research 34, 83-105.

Hadlock, C. and J. Pierce, 2010, "New Evidence on Measuring Financial Constraints: Moving Beyond the K.Z Index," Review of Financial Studies 23, 1909-1940.

Hail, L. and C. Leuz, 2006, "International Differences in Cost of Equity: Do Legal Institutions and Securities Regulation Matter?" Journal of Accounting Research 44, 485-531.

Jensen, M. and W. Meckling, 1976, "Theory of the Firm: Managerial Behavior, Agency Costs, and Capital Structure," Journal of Financial Economics 3, 11-25.

Jones, J., 1991, "Earnings Management During Import Relief Investigations," Journal of Accounting Research 29, 193-228.

Kothari, S., A. Leone, and C. Wasley, 2005, "Performance Matched Discretionary Accrual Measures," Journal of Accounting and Economics 39, 163-197.

Kubick, T., G. Lockhart, and J. Robinson, 2014, "Internal Capital and Investment: Evidence from 2012 Pension Relief," University of Kansas, Clemson University, and University of Texas Austin Working paper.

Leuz, C. and R. Verrecchia, 2005, "Firms' Capital Allocation Choices, Information Quality and the Cost of Capital," University of Pennsylvania Working paper.

Linck, J.S., J. Netter, and T. Shu, 2013, "Can Managers Use Discretionary Accruals to Ease Financial Constraints? Evidence from Discretionary Accruals Prior to Investment," Accounting Review 88, 2117-2143.

Marchica, M. and R. Mura, 2010, "Financial Flexibility, Investment Ability, and Firm Value: Evidence from Firms with Spare Debt Capacity," Financial Management 39, 1339-1365.

Mclnnis, J., 2010, "Earnings Smoothness, Average Returns, and Implied Cost of Equity Capital f Accounting Review 85, 315-341.

Munnell, A. and M. Soto, 2004, "The Outlook for Pension Contributions and Profits," Journal of Pension Economics and Finance 3, 77-97.

Myers, S., 1977, "Determinants of Corporate Borrowing," Journal of Financial Economics 5, 147-175.

Myers, S. and N. Majluf, 1984, "Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have," Journal of Financial Economics 13, 187-221.

Pastor. L.. M. Sinha, and B. Swaminathan, 2008, "Estimating the Inter-temporal Risk-Return Trade-Off Using the Implied Cost of Capital," Journal of Finance 63, 2859-2897.

Phan, H. and S. Hegde, 2013, "Pension Contributions and Firm Performance: Evidence from Frozen Defined Benefit Plans," Financial Management 42, 373-411.

Poterba, J. and L. Summers, 1984, "New Evidence That Taxes Affect the Valuation of Dividends," Journal of Finance 39, 1397-1415.

Pulvino, T. and V Tarhan, 2006, "Cash Flow Sensitivities with Constraints," Northwestern University Working paper.

Rauh, J., 2006, "Investment and Financing Constraints: Evidence from the Funding of Corporate Pension Plans," Journal of Finance 61, 33-71.

Reeb, D., S. Mansi, and J. Alee, 2001, "Firm Internationalization and the Cost of Debt Financing: Evidence from Nonprovisional Publicly Traded Debt," Journal of Financial and Quantitative Analysis 36, 395-414.

Roychowdhury, S., 2006, "Earnings Management Through Real Activities Manipulation," Journal of Accounting and Economics 42, 335-370.

Subramanyam, K., 1996, "The Pricing of Discretionary Accruals," Journal of Accounting and Economics 22, 249-281.

Teoh, S.H., I. Welch, and T.J. Wong, 1998, "Earnings Management and the Long-Run Market Performance of Initial Public Offerings," Journal of Finance 53, 1935-1974.

Trueman, B. and S. Titman, 1988, "An Explanation for Accounting Income Smoothing," Journal of Accounting Research 26, 127-139.

Zang, A., 2012, "Evidence on the Trade-Off Between Real Activities Manipulation and Accrual-Based Earnings Management," Accounting Review 87, 675-703.

(1) We thank the referee for suggesting this grouping of earnings management methods.

(2) Earnings management can also be costly in several ways. It can increase tax expenses, divert management and staff attention to unproductive activities, and bring on costly litigation from regulators and investors. For example, Cohen and Zarowin (2010) find that firms manipulate earnings prior to seasoned equity offerings to help boost offering proceeds, but firms' future operating performance suffers. Teoh, Welch, and Wong (1998) conclude that three-year aftermarket stock returns are 20% lower for IPO issuers that aggressively manage earnings compared to those that are conservative. DuCharme, Malatesta, and Sefcik (2004) note that managers are able to send false signals about their firm's health prior to stock issues, but they later face litigation. McInnis (2010) finds that investors see through earnings smoothing, and that smoothing firms enjoy no decrease in implied costs of equity.

(3) The Pension Protection Act (PPA) of 2006 requires the amortization of funding shortfalls in seven years beginning in 2008. As such, we calculate MFC as the sum of normal costs plus l/7th of the funding gap for 2008 and later, but our results are qualitatively similar.

(4) We follow Campbell et al. (2012) in scaling MC by book assets and funded status by the market value of equity, but our results are essentially unchanged if we scale both MC and funded status by the same measure (i.e., either total assets or the market value of equity).

(5) It is noteworthy that REM, by accelerating sales, can lead to larger actual cash flows although the difference between the predicted and actual cash flows (i.e.. ABN_CFO) is expected to increase.

(6) We note that the magnitude of our coefficient estimate of MC is smaller than that of Campbell et al. (2012), perhaps due to the fact that we use plan-level pension data from Form 5500s to estimate MC, while they use firm-level pension data reported in Compustat for their estimation.

(7) We acknowledge that the new debt issue subsample used in our cost of debt regressions may under represent the whole sample as this subsample includes only those firms that can access the external debt market. However, if earnings management helps firms raise external financing, we expect the effect of earnings management on financing to be more pronounced for this particular subsample.

(8) Campello and Graham (2013) suggest that Compustat data might confound equity issuance programs that are meant to bring in new external funds with other activities, such as the conversion of convertible debt or exercise of stock options.

Thus, in an unreported analysis, we replace the Compustat-based net debt and equity issued with SDC's new debt and equity issuance, but our findings are qualitatively similar.

(9) We obtain audit companies and tenure data from Audit Analytics, but these data are available only from 1999 onward. For a robustness check, similar to Cornett, McNutt, and Tehranian (2009), we use the number of I/B/E/S analysts, the number of audit committee members, and the number of audit committee's annual meetings as instruments for accrual-based earnings management. The number of audit committee members and meetings are hand-collected from the DEF 14A proxy reports that firms file annually with the Securities Exchange Commission (SEC) that are archived in the EDGAR database. Given the large panel of firms, to make our data collection feasible, we focus the analysis on a subsample of bond issuers and use these collected variables as substitutes for auditor quality and auditor tenure in the IV regressions. However, our results are essentially unchanged.

(10) Unlike Zang (2012), we do not use the Altman Z-score and institutional ownership as instruments for real earnings management as these variables do not pass the instrument validity test.

(11) We do not include R&D investment in the aggregated investment when the test variables are ABN_DISX, RM1, and RM2 to avoid a possible mechanical relation between the dependent and the test variables. Alternatively, we additionally account for other forms of investment by aggregating capital expenditures, R&D investment, and acquisitions to firm-year level and re-estimate the investment model, but our findings are qualitatively unchanged.

(12) In an alternative analysis, we follow Campbell et al. (2012) in using service cost, the fair value of pension plan assets (FVPA), and the accumulated benefit obligation (ABO) reported in Compustat to estimate MCs and use them in our analysis, but our findings hold.

(13) We thank an anonymous referee for these suggestions.

(14) We thank an anonymous referee for suggesting the possibility that a firm's plants in different locations may have different EINs as an additional reason for using company name matching.

We appreciate the helpful comments from Raghu Rau (Editor), an anonymous referee, session participants at the 2014 Financial Management Association International Annual Meetings, and participants at a Prairie View A&M University College of Business research seminar. We thank Huong Doan, Dylan Mooney, and Bharadwaj Rachamadugubalaji for excellent research assistance. All remaining errors are ours.

Hieu V. Phan, Hinh D. Khieu, and Joseph Golec *

* Hieu V. Phan is an Assistant Professor in the Department of Finance in the Manning School of Business at the University of Massachusetts Lowell, MA. Hinh D. Khieu is an Assistant Professor of Finance in the Department of Accounting, Finance, and MIS in the College of Business at Prairie View A&M University, TX. Joseph Golec is a Professor in the Department of Finance in the School of Business at the University of Connecticut in Storrs, CT.
Table 1. Descriptive Statistics

The table reports the summary statistics of the main variables of
the sample. Measures of noncash-generating earnings management (NEM)
and cash-generating earnings management (CEM) are described in the
text and Appendix A. The variables PDA1, PDA2, and ABN_PROD are
alternative measures of the NEM. Measures of CEM include ABN_DISX,
ABN_CFO, and RM. We use the plan-level data reported on Form 5500s
to estimate MC as max (MFC, DRC), where MFC is the minimum funding
contribution and DRC is the deficit reduction contribution. MFC is
approximated as the normal cost, which is the present value of
pension benefits accrued during the year, plus 10% of the previous
period's funding gap. DRC as a fraction of the funding gap is
calculated as the min (0.30, [0.30-0.25 (pension assets-pension
liabilities--0.35)]) of the underfunding amount for 1991 to 1994,
and as the min (0.30, [0.30-0.40 (pension assets/pension
liabilities--0.60)]) of the underfunding amount for 1995 and later.
MCs are aggregated from the plan-to firm-level. MC is the ratio of
mandatory pension contributions to book assets at the beginning of
the year. Funded Status is measured as pension assets minus pension
liabilities, both at the plan-level, then aggregated to firm-level
and divided by the market value of equity at the beginning of the
year. Market-to-Book Ratio is the ratio of the market value of
assets to the book value of assets. Book Leverage is the ratio of
the sum of short-term and long-term debts to book assets. The
remaining variables are defined in Appendix B.

                                         N         Mean         p25

Discretionary accrual 1 (PDA1)         12,773      -0.007      -0.055
Discretionary accrual 2 (PDA2)         12,773      -0.007      -0.055
Abnormal sales cash flows from         12,773      -0.017      -0.060
  operations (ABN_CFO)
Abnormal production costs              12,773       0.313      -0.132
  (ABN_PROD)
Abnormal discretionary                  4,897      -0.005      -0.012
expenditures (ABN_DISX)
Real earnings management index          4,897      -0.049      -0.100
  (RM)
Funded Status                          12,773       0.010      -0.008
MC                                     12,773       0.002       0.000
ROA                                    12,773       0.038       0.014
Book Leverage                          12,773       0.267       0.142
Market Value of Equity (in             12,773    7682.962     242.895
  million $)
Market-to-Book Ratio                   12,773       1.662       1.135
Implied Cost of Equity                  8,362       0.092       0.059
Yield Spread                            1,435       0.022       0.008
Net Debt Proceeds (in million $)        1,435     376.787     172.055
Years to Maturity                       1,435      11.540       6.089
Senior Debt Dummy                       1,435       0.947       1.000
Moody Credit Rating                     1,435       9.170       6.000

                                                             Standard
                                      Median        p75      Deviation

Discretionary accrual 1 (PDA1)         -0.004       0.043       0.131
Discretionary accrual 2 (PDA2)         -0.004       0.043       0.132
Abnormal sales cash flows from          0.006       0.040       0.277
  operations (ABN_CFO)
Abnormal production costs               0.045       0.251       2.835
  (ABN_PROD)
Abnormal discretionary                  0.012       0.042       0.168
expenditures (ABN_DISX)
Real earnings management index         -0.031       0.028       0.133
  (RM)
Funded Status                           0.001       0.019       0.125
MC                                      0.000       0.002       0.004
ROA                                     0.047       0.080       0.307
Book Leverage                           0.248       0.359       0.200
Market Value of Equity (in           1098.558    4278.543   25688.650
  million $)
Market-to-Book Ratio                    1.409       1.869       0.962
Implied Cost of Equity                  0.088       0.119       0.043
Yield Spread                            0.015       0.030       0.019
Net Debt Proceeds (in million $)      293.466     494.725     332.522
Years to Maturity                      10.136      10.183      10.724
Senior Debt Dummy                       1.000       1.000       0.225
Moody Credit Rating                     9.000      11.000       4.067

Table II. The Relation between DB Pension Plans and Earnings
Management

The table reports the regression results for the relation between
pension funding status and measures of noncash-generating earnings
management (NEM) or cash-generating earnings management (CEM) in
Panels A and B. respectively. Measures of NEM and CEM are described
in the text and Appendix A. The regression model is Equation (4).
The dependent variables PDA1, PDA2, and ABN_PROD in Panel A are
alternative measures of NEM. Measures of CEM, which include
ABN_DISX, ABN_CFO, and RM, are used as dependent variables in Panel
B. MC is the ratio of mandatory pension contributions to book assets
at the beginning of the year. Funded Status is measured as pension
assets minus pension liabilities, both at the plan-level, then
aggregated to firm-level and divided by the market value of equity
at the beginning of the year. Market-to-Book Ratio is the ratio of
the market value of assets to the book value of assets. Book
Leverage is the ratio of the sum of short-term and long-term debts
to book assets. Book-to-Market Ratio is the ratio of the book value
of equity to the market value of equity. Sales Growth Volatility is
the standard deviation of sales growth during the last five years.
The remaining variables are controls and are defined in Appendix B.
Heteroskedasticity-robust standard errors clustered by firms are
reported in parentheses.

Panel A. DB Pension Plans and Noncash-Generating Earnings
Management

                                             PDA1

                                  (1)            (2)

Lagged Funded Status              -0.027
                                  (0.024)
MC                                -0.366         -0.309
                                  (0.334)        (0.301)
Lagged log(Asset Value)            0.003 ***      0.003 ***
                                  (0.001)        (0.001)
Lagged Market-to-Book Ratio       -0.002         -0.003
                                  (0.002)        (0.002)
Lagged Book Leverage              -0.016 **      -0.016 **
                                  (0.007)        (0.007)
Lagged Sales Growth Volatility     0.009 ***      0.008 ***
                                  (0.003)        (0.003)
Log (Firm Age)                    -0.006 ***     -0.006 **
                                  (0.002)        (0.002)
Intercept                         -0.014         -0.018
                                  (0.013)        (0.015)
Year fixed effects                  Yes            Yes
Industry fixed effects              No             No
Observations                      10,981         12,655
Adjusted [R.sup.2]                 0.01           0.01

                                             PDA2

                                  (3)            (4)

Lagged Funded Status              -0.024
                                  (0.025)
MC                                -0.384         -0.355
                                  (0.334)        (0.301)
Lagged log(Asset Value)            0.003 ***      0.003 ***
                                  (0.001)        (0.001)
Lagged Market-to-Book Ratio       -0.002         -0.003
                                  (0.002)        (0.002)
Lagged Book Leverage              -0.015 **      -0.015 **
                                  (0.007)        (0.007)
Lagged Sales Growth Volatility     0.009 **       0.008 **
                                  (0.003)        (0.003)
Log (Firm Age)                    -0.006 ***     -0.006 ***
                                  (0.002)        (0.002)
Intercept                         -0.016         -0.018
                                  (0.013)        (0.015)
Year fixed effects                  Yes            Yes
Industry fixed effects              No             No
Observations                      10,981         12,655
Adjusted [R.sup.2]                 0.01           0.01

                                           ABN_PROD

                                  (5)            (6)

Lagged Funded Status              -0.341 **
                                  (0.160)
MC                                -5.768 ***     -4.191 **
                                  (1.809)        (1.729)
Lagged log(Asset Value)            0.023 ***      0.025 ***
                                  (0.005)        (0.005)
Lagged Market-to-Book Ratio       -0.076 ***     -0.076 ***
                                  (0.009)        (0.009)
Lagged Book Leverage               0.012          0.039
                                  (0.052)        (0.048)
Lagged Sales Growth Volatility     0.006          0.021
                                  (0.022)        (0.024)
Log (Firm Age)                    -0.005         -0.008
                                  (0.014)        (0.013)
Intercept                          0.036          0.04
                                  (0.063)        (0.059)
Year fixed effects                  Yes            Yes
Industry fixed effects              Yes            Yes
Observations                      10,970         10,970
Adjusted [R.sup.2]                 0.08           0.09

Panel B. DB Pension Plans and Cash-Generating Earnings Management

                                   ABN_DISX     ABN_CFO        RM

                                     (1)          (2)          (3)

Lagged Funded Status              -0.098 **     0.029       -0.838 *
                                  (0.039)      (0.044)      (0.450)
MC                                 1.498 ***    1.621 **    19.156 **
                                  (0.539)      (0.637)      (7.734)
Lagged log (Asset Value)          -0.002       -0.008 ***   -0.027
                                  (0.002)      (0.002)      (0.028)
Lagged Market-to-Book Ratio       -0.012 ***   -0.022 ***   -0.088 **
                                  (0.003)      (0.003)      (0.044)
Lagged Book Leverage               0.013        0.034 **    -0.055
                                  (0.017)      (0.014)      (0.122)
Lagged Sales Growth Volatility     0.022        0.004        0.280 *
                                  (0.023)      (0.005)      (0.151)
Log (Firm Age)                     0.005        0.002       -0.049
                                  (0.004)      (0.004)      (0.061)
Intercept                          0.008        0.068 **     0.11
                                  (0.019)      (0.035)      (0.164)
Industry fixed effects              Yes          Yes          Yes
Year fixed effects                  Yes          Yes          Yes
Observations                       4,044       10,981        4,044
Adjusted [R.sup.2]                 0.05         0.04         0.05

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table III. The Relation between DB Pension Plans and Earnings
Management: Investment Versus Speculative Grade Credit Ratings

The table reports the regression results for the relation between
pension funding status and measures of cash-generating earnings
management (CEM) for subgroups of firms sorted by S&P long-term
credit ratings. Measures of CEM, which include ABN_DISX, ABN_CFO,
and RM, are used as dependent variables. MC is the ratio of
mandatory pension contributions to book assets at the beginning of
the year. Funded Status is measured as pension assets minus pension
liabilities, both at the plan-level, then aggregated to firm-level
and divided by the market value of equity at the beginning of the
year. Market-to-Book Ratio is the ratio of the market value of
assets to the book value of assets. Book Leverage is the ratio of
the sum of short-term and long-term debts to book assets.
Book-to-Market Ratio is the ratio of the book value of equity to the
market value of equity. Sales Growth Volatility is the standard
deviation of sales growth during the last five years. Investment
Grade includes S&P long-term credit ratings of BBB or better.
Speculative Grade includes S&P long-term credit ratings below BBB.
The remaining variables are controls and are defined in Appendix B.
Heteroskedasticity-robust standard errors clustered by firms are
reported in parentheses.

                                          ABN_DISX

                                   Investment    Speculative
                                     Grade          Grade

MC                                 0.264          1.636 *
                                  (0.619)        (0.878)
Lagged Funded Status              -0.105         -0.167 **
                                  (0.065)        (0.085)
Lagged log(Asset Value)            0.004          0.006
                                  (0.004)        (0.006)
Lagged Market-to-Book Ratio       -0.003         -0.021 ***
                                  (0.004)        (0.008)
Lagged Book Leverage               0.088 **       0.007
                                  (0.035)        (0.035)
Lagged Sales Growth Volatility     0.006         -0.030
                                  (0.024)        (0.069)
Log (Firm Age)                    -0.010          0.003
                                  (0.008)        (0.006)
Intercept                         -0.036          0.019
                                  (0.052)        (0.043)
Industry fixed effects              Yes            Yes
Year fixed effects                  Yes            Yes
Observations                       1,630           782
Adjusted [R.sup.2]                 0.05           0.01
Comparing coefficients on MC:
[chi square]                       3.01
p-Value                            0.08

                                           ABN_CFO

                                   Investment    Speculative
                                     Grade          Grade

MC                                -0.715          2.026 **
                                  (1.096)        (0.825)
Lagged Funded Status               0.068         -0.055
                                  (0.088)        (0.080)
Lagged log(Asset Value)            0.001         -0.016 ***
                                  (0.005)        (0.006)
Lagged Market-to-Book Ratio       -0.018 ***     -0.025 ***
                                  (0.005)        (0.008)
Lagged Book Leverage               0.030          0.035
                                  (0.034)        (0.025)
Lagged Sales Growth Volatility     0.014          0.015
                                  (0.018)        (0.031)
Log (Firm Age)                    -0.001          0.006
                                  (0.011)        (0.007)
Intercept                         -0.021          0.127 **
                                  (0.039)        (0.049)
Industry fixed effects              Yes            Yes
Year fixed effects                  Yes            Yes
Observations                       3,593          2,726
Adjusted [R.sup.2]                 0.05           0.04
Comparing coefficients on MC:
[chi square]                       4.12
p-Value                            0.04

                                              RM

                                   Investment    Speculative
                                     Grade          Grade

MC                                 9.212          24.795 **
                                  (9.440)        (11.598)
Lagged Funded Status              -0.208          -2.976 *
                                  (0.365)         (1.768)
Lagged log(Asset Value)            0.037           0.097
                                  (0.076)         (0.128)
Lagged Market-to-Book Ratio       -0.081          -0.155 *
                                  (0.078)         (0.090)
Lagged Book Leverage              -0.414           0.582
                                  (0.446)         (0.410)
Lagged Sales Growth Volatility    -0.008           0.401
                                  (0.287)         (0.361)
Log (Firm Age)                    -0.231 *        -0.257 *
                                  (0.125)         (0.147)
Intercept                          0.686           0.201
                                  (0.744)         (0.911)
Industry fixed effects              Yes            Yes
Year fixed effects                  Yes            Yes
Observations                       1,630           782
Adjusted [R.sup.2]                 0.05           0.01
Comparing coefficients on MC:
[chi square]                       2.74
p-Value                            0.10

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table IV. The Relation between Earnings Management and Costs of Debt

The table reports the regression results for the relation between
the costs of debt and noncash-generating earnings management (NEM)
and cash-generating earnings management (CEM) in Panels A and B,
respectively. The regression model is Equation (5). The dependent
variable, cost of debt, is measured as the difference between yields
on the firm's bond issue and a US Treasury issue of similar
maturity. Measures of NEM and CEM are described in the text and
Appendix A. The independent variables PDA1, PDA2, and ABN_PROD in
Panel A are alternative measures of NEM. Measures of CEM, which
include ABN_DISX, ABN_CFO, and RM, are used as independent variables
in Panel B. MC is the ratio of mandatory pension contributions to
book assets at the beginning of the year. Funded Status is measured
as pension assets minus pension liabilities, both at the plan-
level, then aggregated to firm-level and divided by the market value
of equity at the beginning of the year. Tobin's Q is measured as the
market value of assets minus deferred taxes, all divided by the book
value of assets. Book Leverage is the ratio of the sum of short-
term and long-term debts to book assets. Return on Assets (ROA) is
measured as income before extraordinary items divided by book assets
at the beginning of the year. Senior Debt Dummy is an indicator
variable that is equal to one if the debt issue is senior and zero
otherwise. Moody Rating Score is the numerical equivalent of the
Moody credit rating where one is assigned to bonds with an Aaa
rating and 19 is assigned to bonds with a C rating. All models are
estimated with industry and year dummies, but those estimates are
not reported. Some control variables are suppressed from Panel B for
brevity. Heteroskedasticity-robust standard errors clustered by
firms are reported in parentheses.

Panel A. Noncash-Generating Earnings Management and Costs of Debt

                           (1)            (2)            (3)

PDA1                                 -0.001         -0.006
                                     (0.003)        (0.004)
PDA1 x MC                                            2.471 ***
                                                    (0.748)
PDA2

PDA2 x MC

ABN_PROD

ABN_PROD x MC

MC                     0.158 *        0.159 *        0.239 **
                      (0.091)        (0.092)        (0.094)
Funded Status         -0.009         -0.009         -0.005
                      (0.010)        (0.010)        (0.010)
Book Leverage          0.007 **       0.007 **       0.007 **
                      (0.003)        (0.003)        (0.003)
Tobin's Q              0.001          0.001          0.001
                      (0.001)        (0.001)        (0.001)
Senior Debt           -0.005 ***     -0.005 ***     -0.005 ***
  Dummy               (0.002)        (0.002)        (0.002)
Log (Bond              0.001          0.001          0.001
  Maturity)           (0.001)        (0.001)        (0.001)
Log (Bond              0.003 ***      0.003 ***      0.003 ***
  Proceeds)           (0.001)        (0.001)        (0.001)
Moody Rating           0.001 ***      0.001 ***      0.001 ***
  Scores              (0.000)        (0.000)        (0.000)
Log(Market Value      -0.004 ***     -0.004 ***     -0.004 ***
  of Equity)          (0.001)        (0.001)        (0.001)
ROA                   -0.044" *      -0.044 ***     -0.049 ***
                      (0.009)        (0.009)        (0.008)
Intercept              0.050 ***      0.050 **       0.038 ***
                      (0.007)        (0.007)        (0.008)
Industry fixed          Yes            Yes            Yes
  effects
Year fixed effects      Yes            Yes            Yes
Number of              1,276          1,276          1,276
  observations
Adjusted [R.sup.2]     0.62           0.62           0.63

                          (4)            (5)

PDA1

PDA1 x MC

PDA2                  -0.002         -0.006
                      (0.003)        (0.005)
PDA2 x MC                             2.231 ***
                                     (0.757)
ABN_PROD

ABN_PROD x MC

MC                     0.159 *        0.214 **
                      (0.092)        (0.095)
Funded Status         -0.009         -0.006
                      (0.010)        (0.010)
Book Leverage          0.007 **       0.007 **
                      (0.003)        (0.003)
Tobin's Q              0.001          0.001
                      (0.001)        (0.001)
Senior Debt           -0.005 ***     -0.005 ***
  Dummy               (0.002)        (0.002)
Log (Bond              0.001          0.001
  Maturity)           (0.001)        (0.001)
Log (Bond              0.003 ***      0.003 ***
  Proceeds)           (0.001)        (0.001)
Moody Rating           0.001 ***      0.001 ***
  Scores              (0.000)        (0.000)
Log(Market Value      -0.004 ***     -0.004 ***
  of Equity)          (0.001)        (0.001)
ROA                   _0.044 ***     -0.049 ***
                      (0.009)        (0.008)
Intercept              0.050 ***      0.038 ***
                      (0.007)        (0.008)
Industry fixed          Yes            Yes
  effects
Year fixed effects      Yes            Yes
Number of              1,276          1,276
  observations
Adjusted [R.sup.2]     0.62           0.62

                          (6)            (7)

PDA1

PDA1 x MC

PDA2

PDA2 x MC

ABN_PROD              -0.001         -0.001
                      (0.001)        (0.001)
ABN_PROD x MC                        -0.045
                                     (0.075)
MC                     0.126          0.129 *
                      (0.077)        (0.078)
Funded Status         -0.010         -0.010
                      (0.010)        (0.010)
Book Leverage          0.007 **       0.007 **
                      (0.003)        (0.003)
Tobin's Q              0.001          0.001
                      (0.001)        (0.001)
Senior Debt           -0.005 ***     -0.005 ***
  Dummy               (0.002)        (0.002)
Log (Bond              0.001          0.001
  Maturity)           (0.001)        (0.001)
Log (Bond              0.003 ***      0.003 ***
  Proceeds)           (0.001)        (0.001)
Moody Rating           0.001 ***      0.001 ***
  Scores              (0.000)        (0.000)
Log(Market Value      -0.004 ***     -0.004 ***
  of Equity)          (0.001)        (0.001)
ROA                   -0.044 ***     -0.044 ***
                      (0.009)        (0.009)
Intercept              0.025 ***      0.042 ***
                      (0.005)        (0.008)
Industry fixed          Yes            Yes
  effects
Year fixed effects      Yes            Yes
Number of              1,276          1,276
  observations
Adjusted [R.sup.2]     0.63           0.63

Panel B. Cash-Generating Earnings Management and Costs of Debt

                              (1)            (2)            (3)

ABN_DISX                   0.001          0.001
                          (0.001)        (0.001)
ABN_DISX x MC                             0.044
                                         (0.112)
ABN_CFO                                                  0.001
                                                        (0.001)
ABN_CFO x MC

RM2

RM2 x MC

MC                         0.096          0.098          0.159 *
                          (0.132)        (0.132)        (0.092)
Funded Status             -0.024 **      -0.024 **      -0.009
                          (0.012)        (0.012)        (0.010)
Other controls              Yes            Yes            Yes
Industry fixed effects      Yes            Yes            Yes
Year fixed effects          Yes            Yes            Yes
Number of observations      578            578           1,276
Adjusted [R.sup.2]         0.60           0.60           0.62

                              (4)            (5)            (6)

ABN_DISX

ABN_DISX x MC

ABN_CFO                   -0.001
                          (0.001)
ABN_CFO x MC               0.205 **
                          (0.104)
RM2                                       0.001          0.001
                                         (0.001)        (0.001)
RM2 x MC                                                 0.236 *
                                                        (0.135)
MC                         0.163 *        0.096          0.121
                          (0.092)        (0.132)        (0.131)
Funded Status             -0.009         -0.024 **      -0.024 **
                          (0.010)        (0.012)        (0.012)
Other controls              Yes            Yes            Yes
Industry fixed effects      Yes            Yes            Yes
Year fixed effects          Yes            Yes            Yes
Number of observations     1,276           578            578
Adjusted [R.sup.2]         0.62           0.60           0.60

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table V. The Relation between Earnings Management and Implied Costs
of Equity

The table reports the regression results for the relation between
the implied cost of equity and noncash-generating earnings
management (NEM) and cash-generating earnings management (CEM) in
Panels A and B, respectively. The regression model is Equation (6).
The dependent variable, implied cost of equity, is estimated as the
internal rate of return that equates the current stock price with
the present value of all future cash flows to common shareholders.
Measures of NEM and CEM are described in the text and Appendix A.
The independent variables PDA1, PDA2, and ABN_PROD in Panel A are
alternative measures of NEM. Measures of CEM, which include
ABN_DISX, ABN_CFO, and RM, are used as independent variables in
Panel B. MC is the ratio of mandatory pension contributions to book
assets at the beginning of the year. Funded Status is measured as
pension assets minus pension liabilities, both at the plan-level,
then aggregated to firm-level and divided by the market value of
equity at the beginning of the year. Book-to-Market Ratio is the
ratio of the book value of equity to the market value of equity.
Book Leverage is the ratio of the sum of short-term and long-term
debts to book assets. Tobin s Q is measured as the market value of
assets minus deferred taxes, all divided by the book value of
assets. Market beta is the capital market beta estimated with the
market model using value-weighted CRSP returns and a minimum of 24
monthly returns over the prior 60 months. Long-Term Growth Rate is
the analysts' long-term growth forecast available from the I-B-E-S
database. Analyst Forecast Dispersion is the natural logarithm of
the standard deviation of analyst estimates for the next period's
earnings divided by the consensus forecast for next period's
earnings. Industry' Cost of Equity is the average cost of equity in
each year for each industry based on the Fama-French (1997) 48
industry classifications. All models are estimated with year
dummies, but those estimates are not reported. Some control
variables are suppressed from Panel B for brevity.
Heteroskedasticity-robust standard errors clustered by firms are
reported in parentheses.

Panel A. Noncash-Generating Earnings Management and Implied Costs of
Equity

                                   (1)           (2)           (3)

PDA1                                          0.008 **      0.011 *
                                             (0.004)       (0.004)
PDA1 x MC                                                  -1.853
                                                           (1.339)
PDA2

PDA2 x MC

ABN_PROD

ABN_PROD x MC

MC                              0.292 **      0.301 **      0.270 *
                               (0.141)       (0.141)       (0.142)
Funded Status                   0.012         0.012         0.012
                               (0.010)       (0.010)       (0.010)
Market Beta                     0.002 *       0.002 **      0.002 **
                               (0.001)       (0.001)       (0.001)
Log(Market Value of Equity)    -0.004 *      -0.004 *      -0.004 ***
                               (0.001)       (0.001)       (0.001)
Book-to-Market Ratio            0.023 ***     0.023 ***     0.023 **
                               (0.003)       (0.003)       (0.003)
  Book Leverage                 0.030 ***     0.030 ***     0.030 **
                               (0.004)       (0.004)       (0.004)
Long-Term Growth Rate           0.066 *       0.065 ***     0.065 **
                               (0.011)       (0.011)       (0.011)
Analyst Forecast Dispersion     0.001         0.001         0.001
                               (0.001)       (0.001)       (0.001)
Industry Cost of Equity         0.132 ***     0.132 *       0.132 ***
                               (0.014)       (0.014)       (0.014)
Intercept                       0.097 **      0.097 *       0.098 *
                               (0.005)       (0.005)       (0.005)
Year fixed effects               Yes           Yes           Yes
Number of observations          5,707         5,707         5,707
Adjusted [R.sup.2]              0.22          0.22          0.22

                                   (4)           (5)

PDA1

PDA1 x MC

PDA2                            0.008 **      0.011 **
                               (0.004)       (0.004)
PDA2 x MC                                    -1.509
                                             (1.161)
ABN_PROD

ABN_PROD x MC

MC                              0.302 **      0.274 *
                               (0.141)       (0.143)
Funded Status                   0.012         0.012
                               (0.010)       (0.010)
Market Beta                     0.002 **      0.002 **
                               (0.001)       (0.001)
Log(Market Value of Equity)    -0.004 *      -0.004 ***
                               (0.001)       (0.001)
Book-to-Market Ratio            0.023 ***     0.023 *
                               (0.003)       (0.003)
  Book Leverage                 0.030 *       0.030 ***
                               (0.004)       (0.004)
Long-Term Growth Rate           0.065 ***     0.065 *
                               (0.011)       (0.011)
Analyst Forecast Dispersion     0.001         0.001
                               (0.001)       (0.001)
Industry Cost of Equity         0.132 ***     0.132 ***
                               (0.014)       (0.014)
Intercept                       0.098 *       0.098 *
                               (0.005)       (0.005)
Year fixed effects               Yes           Yes
Number of observations          5,707         5,707
Adjusted [R.sup.2]              0.22          0.22

                                   (6)           (7)

PDA1

PDA1 x MC

PDA2

PDA2 x MC

ABN_PROD                        0.001 *       0.001
                                0.000        (0.001)
ABN_PROD x MC                                -0.001
                                             (0.021)
MC                              0.294 **      0.294 **
                               (0.141)       (0.141)
Funded Status                   0.012         0.012
                               (0.010)       (0.010)
Market Beta                     0.002 **      0.002 **
                               (0.001)       (0.001)
Log(Market Value of Equity)    -0.004 **     -0.004 ***
                               (0.001)       (0.001)
Book-to-Market Ratio            0.023 ***     0.023 ***
                               (0.003)       (0.003)
  Book Leverage                 0.030 **      0.030 ***
                               (0.004)       (0.004)
Long-Term Growth Rate           0.065 ***     0.065 *
                               (0.011)       (0.011)
Analyst Forecast Dispersion     0.001         0.001
                               (0.001)       (0.001)
Industry Cost of Equity         0.132 *       0.132 **
                               (0.014)       (0.014)
Intercept                       0.097 *       0.097 ***
                               (0.005)       (0.005)
Year fixed effects               Yes           Yes
Number of observations          5,707         5,707
Adjusted [R.sup.2]              0.22          0.22

Panel B. Cash-Generating Earnings Management and Implied Costs of
Equity

                              (1)           (2)           (3)

ABN_DISX                   0.001 ***     0.001 ***
                          (0.000)       (0.000)
ABN_DISX x MC                           -0.738 ***
                                        (0.273)
ABN_CFO                                                0.001
                                                      (0.001)
ABN_CFO x MC

RM

RM x MC

MC                         0.410 **      0.394 **      0.294 **
                          (0.187)       (0.184)       (0.141)
Funded Status              0.024 *       0.023 *       0.012
                          (0.013)       (0.013)       (0.010)
Other controls              Yes           Yes           Yes
Year fixed effects          Yes           Yes           Yes
Number of observations     2,649         2,649         5,707
Adjusted [R.sup.2]         0.17          0.17          0.22

                              (4)           (5)           (6)

ABN_DISX

ABN_DISX x MC

ABN_CFO                    0.001
                          (0.001)
ABN_CFO x MC              -0.172
                          (0.294)
RM                                       0.001         0.001
                                        (0.001)       (0.001)
RM x MC                                               -0.228
                                                      (0.249)
MC                         0.287 **      0.412 **      0.394 **
                          (0.140)       (0.187)       (0.185)
Funded Status              0.011         0.023 *       0.022 *
                          (0.010)       (0.013)       (0.013)
Other controls              Yes           Yes           Yes
Year fixed effects          Yes           Yes           Yes
Number of observations     5,707         2,649         2,649
Adjusted [R.sup.2]         0.22          0.17          0.17

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table VI. The Relation between Earnings Management and the Weighted
Average Cost of Capital (WACC)

The table reports the regression results for the relation between
WACC and noncash-generating earnings management (NEM) and cash-
generating earnings management (CEM) in Panels A and B,
respectively. The dependent variable, WACC, is the weighted average
of a firm's implied cost of equity and its after-tax cost of debt.
Measures of NEM and CEM are described in the text and Appendix A.
The dependent variables PDA1, PDA2, and ABN_PROD in Panel A are
alternative measures of NEM. Measures of CEM, which include
ABN_DISX, ABN_CFO, and RM, are used as dependent variables in Panel
B. MC is the ratio of mandatory pension contributions to book assets
at the beginning of the year. Funded Status is measured as pension
assets minus pension liabilities, both at the plan-level, then
aggregated to firm-level and divided by the market value of equity
at the beginning of the year. Market-to-Book Ratio is the ratio of
the market value of assets to the book value of assets. Book
Leverage is the ratio of the sum of short-term and long-term debts
to book assets. Book-to-Market Ratio is the ratio of the book value
of equity to the market value of equity. Sales Growth Volatility is
the standard deviation of sales growth during the last five years.
The remaining variables are controls and are defined in Appendix B.
All models are estimated with year dummies, but those estimates are
not reported. Some control variables are suppressed from Panel B for
brevity. Heteroskedasticity-robust standard errors clustered by
firms are reported in parentheses.

Panel A. Noncash-Generating Earnings Management and WACC

                                   (1)           (2)           (3)

PDA1                                          0.006         0.011
                                             (0.007)       (0.008)
PDA1 x MC                                                  -2.71
                                                           (1.768)
PDA2

PDA2 x MC

ABN_PROD

ABN_PROD x MC

MC                              0.234 *       0.077         0.097
                               (0.139)       (0.137)       (0.142)
Funded Status                   0.031         0.020         0.021
                               (0.021)       (0.021)       (0.021)
Market Beta                     0.003 **      0.004 ***     0.004 ***
                               (0.002)       (0.002)       (0.002)
Log(Market Value of Equity)    -0.002 **     -0.002 **     -0.002 ***
                               (0.001)       (0.001)       (0.001)
Book-to-Market Ratio            0.005 *       0.005 *       0.005 *
                               (0.003)       (0.003)       (0.003)
Book Leverage                  -0.011        -0.013 **     -0.013 **
                               (0.007)       (0.007)       (0.007)
Long-Term Growth Rate           0.055 **      0.043 **      0.042 **
                               (0.018)       (0.018)       (0.018)
Analyst Forecast Dispersion     0.001         0.001         0.001
                               (0.001)       (0.001)       (0.001)
Industry Cost of Equity         0.055 **      0.005         0.007
                               (0.024)       (0.036)       (0.036)
Intercept                       0.088 ***     0.087 ***     0.087 ***
                               (0.007)       (0.007)       (0.007)
Year fixed effects               Yes           Yes           Yes
Number of observations           983           983           983
Adjusted [R.sup.2]              0.2           0.25          0.25

                                   (4)           (5)

PDA1

PDA1 x MC

PDA2                            0.007         0.011
                               (0.007)       (0.008)
PDA2 x MC                                    -2.358
                                             (1.719)
ABN_PROD

ABN_PROD x MC

MC                              0.078         0.092
                               (0.147)       (0.141)
Funded Status                   0.021         0.021
                               (0.021)       (0.021)
Market Beta                     0.004 ***     0.004 ***
                               (0.002)       (0.002)
Log(Market Value of Equity)    -0.002 **     -0.002 **
                               (0.001)       (0.001)
Book-to-Market Ratio            0.005 *       0.005 *
                               (0.003)       (0.003)
Book Leverage                  -0.013 **     -0.013 **
                               (0.007)       (0.007)
Long-Term Growth Rate           0.042 **      0.042 **
                               (0.018)       (0.018)
Analyst Forecast Dispersion     0.001         0.001
                               (0.001)       (0.001)
Industry Cost of Equity         0.005         0.007
                               (0.036)       (0.036)
Intercept                       0.087 ***     0.088 ***
                               (0.007)       (0.007)
Year fixed effects               Yes           Yes
Number of observations           983           983
Adjusted [R.sup.2]              0.25          0.25

                                   (6)           (7)

PDA1

PDA1 x MC

PDA2

PDA2 x MC

ABN_PROD                        0.001         0.001
                               (0.001)       (0.001)
ABN_PROD x MC                                 0.129
                                             (0.114)
MC                              0.235         0.214
                               (0.145)       (0.144)
Funded Status                   0.031         0.031
                               (0.021)       (0.021)
Market Beta                     0.003 **      0.003 **
                               (0.002)       (0.002)
Log(Market Value of Equity)    -0.002 ***    -0.002 ***
                               (0.001)       (0.001)
Book-to-Market Ratio            0.005 *       0.005 *
                               (0.003)       (0.003)
Book Leverage                  -0.011        -0.011
                               (0.007)       (0.007)
Long-Term Growth Rate           0.055 **      0.056 *
                               (0.018)       (0.018)
Analyst Forecast Dispersion     0.001         0.001
                               (0.001)       (0.001)
Industry Cost of Equity         0.055 **      0.055 **
                               (0.024)       (0.024)
Intercept                       0.088 *       0.088 *
                               (0.007)       (0.007)
Year fixed effects               Yes           Yes
Number of observations           983           983
Adjusted [R.sup.2]              0.19          0.19

Panel B. Cash-Generating Earnings Management and WACC

                                   (1)           (2)           (3)

ABN_DISX                        0.001         0.001
                               (0.001)       (0.001)
ABN_DISX x MC                                -0.353 ***
                                             (0.102)
ABN_CFO                                                     0.001
                                                           (0.001)
ABN_CFO x MC

RM

RM x MC

MC                              0.413 *       0.393 *       0.235 *
                               (0.236)       (0.237)       (0.137)
Funded Status                   0.025         0.026         0.031
                               (0.027)       (0.027)       (0.021)
Market Beta                    -0.004        -0.004         0.003 **
                               (0.003)       (0.003)       (0.002)
Other controls                   Yes           Yes           Yes
Year fixed effects               Yes           Yes           Yes
Number of observations           453           453           983
Adjusted [R.sup.2]              0.13          0.14          0.19

                                   (4)           (5)           (6)

ABN_DISX

ABN_DISX x MC

ABN_CFO                         0.001
                               (0.001)
ABN_CFO x MC                   -0.367
                               (0.539)
RM                                            0.001         0.001
                                             (0.001)       (0.001)
RM x MC                                                    -0.312 ***
                                                           (0.061)
MC                              0.232 *       0.415 *       0.378 *
                               (0.133)       (0.228)       (0.226)
Funded Status                   0.031         0.025         0.026
                               (0.021)       (0.027)       (0.027)
Market Beta                     0.003 **     -0.004        -0.004
                               (0.002)       (0.003)       (0.003)
Other controls                   Yes           Yes           Yes
Year fixed effects               Yes           Yes           Yes
Number of observations           983           453           453
Adjusted [R.sup.2]              0.19          0.13          0.13

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table VII. The Relation between Earnings Management and External
Capital Issuances

The table reports the regression results for the relation between
capital issuance and noncash-generating earnings management (NEM)
and cash-generating earnings management (CEM). The dependent
variables are net debt issuance (D) or net equity issuance (E). Net
debt issuance is measured as the difference between long-term debt
issuance and long-term debt reduction, all divided by book assets at
the beginning of the year. Net equity issuance is measured as the
difference between the sales of common stock and stock repurchases,
all divided by book assets at the beginning of the year. Measures of
NEM and CEM are described in the text and Appendix A. The
independent variables PDA1, PDA2, and ABN_PROD are alternative
measures of NEM. Measures of CEM include ABN_DISX, ABN_CFO, and RM.
MC is the ratio of mandatory pension contributions to book assets at
the beginning of the year. Funded Status is measured as pension
assets minus pension liabilities, both at the plan-level, then
aggregated to firm-level and divided by the market value of equity
at the beginning of the year. Book Leverage is the ratio of the sum
of short-term and long-term debt to book assets. Tobin's Q is
measured as the market value of assets minus deferred taxes, all
divided by the book value of assets. Cashflow is the net cash flow
from operating activities divided by book assets at the beginning of
the year. Dividend is measured as the sum of dividends to common and
preferred stocks divided by book assets at the beginning of the
year. All models are estimated with year dummies, but those
estimates are not reported. Some control variables are suppressed
from Panel B. Heteroskedasticity-robust standard errors clustered by
firms are reported in parentheses.

Panel A. Earnings Management and Debt Issuances

                                   (1)           (2)           (3)

Lagged PDA1                     0.008
                               (0.010)
Lagged PDA1 x lagged MC        -0.926
                               (1.860)
Lagged PDA2                                   0.007
                                             (0.010)
Lagged PDA2 x lagged MC                      -0.951
                                             (1.928)
Lagged ABN_PROD                                             0.001
                                                           (0.001)
Lagged ABN_PROD x lagged MC                                 0.001
                                                           (0.039)
Lagged ABN_DISX

Lagged ABN_DISX x lagged MC

Lagged ABN_CFO

Lagged ABN_CFO x lagged MC

Lagged RM

Lagged RM x lagged MC

Lagged MC                       0.012         0.010         0.011
                               (0.332)       (0.331)       (0.334)
Lagged Funded Status           -0.001        -0.001        -0.001
                               (0.023)       (0.023)       (0.023)
Cash Flow                       0.020         0.020         0.024
                               (0.022)       (0.022)       (0.022)
Book Leverage                  -0.063 ***    -0.063 ***    -0.062 ***
                               (0.007)       (0.007)       (0.007)
Tobin's Q                       0.009 ***     0.009 ***     0.008 ***
                               (0.002)       (0.002)       (0.002)
Dividend                       -0.053        -0.053        -0.048
                               (0.046)       (0.046)       (0.044)
Cash Ratio                     -0.026 **     -0.026 **     -0.024 **
                               (0.010)       (0.010)       (0.010)
Sales Growth                    0.027 ***     0.027 ***     0.027 ***
                               (0.007)       (0.007)       (0.007)
Size                           -0.001        -0.001        -0.001
                               (0.001)       (0.001)       (0.001)
Intercept                       0.019 ***     0.019 ***     0.019 ***
                               (0.007)       (0.007)       (0.007)
Year fixed effects               Yes           Yes           Yes
Number of observations         10,301        10,301        10,301
Adjusted [R.sup.2]              0.03          0.03          0.03

                                   (4)           (5)           (6)

Lagged PDA1

Lagged PDA1 x lagged MC

Lagged PDA2

Lagged PDA2 x lagged MC

Lagged ABN_PROD

Lagged ABN_PROD x lagged MC

Lagged ABN_DISX                -0.001
                               (0.001)
Lagged ABN_DISX x lagged MC     0.48
                               (0.864)
Lagged ABN_CFO                                0.001
                                             (0.001)
Lagged ABN_CFO x lagged MC                   -0.518
                                             (0.467)
Lagged RM                                                   0.001
                                                           (0.001)
Lagged RM x lagged MC                                      -0.204
                                                           (0.579)
Lagged MC                       0.190         0.009         0.165
                               (0.574)       (0.332)       (0.576)
Lagged Funded Status            0.003        -0.001         0.002
                               (0.035)       (0.023)       (0.035)
Cash Flow                      -0.053         0.019        -0.054
                               (0.047)       (0.022)       (0.047)
Book Leverage                  -0.064 ***    -0.064 ***    -0.064 ***
                               (0.013)       (0.007)       (0.013)
Tobin's Q                       0.009 **      0.009 ***     0.009 **
                               (0.004)       (0.002)       (0.004)
Dividend                        0.060        -0.053         0.060
                               (0.075)       (0.046)       (0.076)
Cash Ratio                     -0.028 *      -0.026"       -0.028 *
                               (0.017)       (0.010)       (0.017)
Sales Growth                    0.028 ***     0.028 **      0.028 ***
                               (0.010)       (0.007)       (0.010)
Size                           -0.001        -0.001        -0.001
                               (0.001)       (0.001)       (0.001)
Intercept                       0.028 *       0.019 ***     0.028 *
                               (0.016)       (0.007)       (0.016)
Year fixed effects               Yes           Yes           Yes
Number of observations          4,155        10,301         4,155
Adjusted [R.sup.2]              0.03          0.03          0.03

Panel B. Earnings Management and Equity Issuances

                                   (1)           (2)           (3)

Lagged PDA1                     0.001
                               (0.004)
Lagged PDA1 x lagged MC        -1.268 **
                               (0.497)
Lagged PDA2                                   0.001
                                             (0.004)
Lagged PDA2 x lagged MC                      -1.121 **
                                             (0.490)
Lagged ABN_PROD                                            -0.001
                                                           (0.001)
Lagged ABN_PROD x lagged MC                                 0.003
                                                           (0.009)
Lagged ABN_DISX

Lagged ABN_DISX x lagged MC

Lagged ABN_CFO

Lagged ABN_CFO x lagged MC

Lagged RM

Lagged RM x lagged MC

Lagged MC                      -0.158 *      -0.157 *      -0.139
                               (0.090)       (0.090)       (0.091)
Lagged Funded Status           -0.004        -0.004        -0.004
                               (0.008)       (0.008)       (0.008)
Other controls                   Yes           Yes           Yes
Year fixed effects               Yes           Yes           Yes
Number of observations         10,699        10,699        10,680
Adjusted [R.sup.2]              0.03          0.03          0.03

                                   (4)           (5)           (6)

Lagged PDA1

Lagged PDA1 x lagged MC

Lagged PDA2

Lagged PDA2 x lagged MC

Lagged ABN_PROD

Lagged ABN_PROD x lagged MC

Lagged ABN_DISX                 0.001
                               (0.001)
Lagged ABN_DISX x lagged MC    -0.025
                               (0.151)
Lagged ABN_CFO                                0.001
                                             (0.001)
Lagged ABN_CFO x lagged MC                   -0.033
                                             (0.098)
Lagged RM                                                   0.001
                                                           (0.001)
Lagged RM x lagged MC                                      -0.013
                                                           (0.132)
Lagged MC                       0.063        -0.143         0.063
                               (0.165)       (0.091)       (0.165)
Lagged Funded Status           -0.011        -0.004        -0.011
                               (0.010)       (0.008)       (0.010)
Other controls                   Yes           Yes           Yes
Year fixed effects               Yes           Yes           Yes
Number of observations          4,346        10,699         4,346
Adjusted [R.sup.2]              0.03          0.03          0.03

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table VIII. The Relation between Earnings Management and Corporate
Investment

The table reports the regression results for the relation between
investment and noncash-generating earnings management (NEM) and
cash-generating earnings management (CEM). The regression model is
Equation (9). The dependent variable, corporate investment, is
measured as capital expenditures divided by book assets at the
beginning of the year. Measures of NEM and CEM are described in the
text and Appendix A. The independent variables PDA1, PDA2, and
ABN_PROD are alternative measures of NEM. CEM measures include
ABN_DISX, ABN_CFO, and RM. MC is the ratio of mandatory pension
contributions to book assets at the beginning of the year. Funded
Status is measured as pension assets minus pension liabilities, both
at the plan-level, then aggregated to firm-level and divided by the
market value of equity at the beginning of the year. Tobin's Q is
measured as the market value of assets minus deferred taxes, all
divided by the book value of assets. Nonpension Cash Flow is
measured as (Net Income + Depreciation and Amortization + Pension
Expense)-Book Assets at the beginning of the year. All models are
estimated with firm fixed effects, but those estimates are not
reported. Panel A presents OLS estimation results with firm-fixed
effects. Panel B provides the GMM estimation results with firm fixed
effects. Heteroskedasticity-robust standard errors clustered by
firms are reported in parentheses.

Panel A. Earnings Management and Corporate Investment--FE Estimation

                                 (1)            (2)           (3)

PDA1                           -0.002
                               (0.002)
PDA1 x MC                      -0.121
                               (0.549)
PDA2                                         -0.002
                                             (0.002)
PDA2 x MC                                    -0.148
                                             (0.533)
ABN_PROD                                                    0.001
                                                           (0.001)
ABN_PROD x MC                                              -0.007
                                                           (0.012)
ABN_DISX

ABN_DISX x MC

ABN CFO

ABN_CFO x MC

RM

RM x MC

MC                             -0.236 *      -0.236 *      -0.233 *
                               (0.127)       (0.127)       (0.129)
Lagged Funded Status            0.068 ***     0.068 **      0.068 ***
                               (0.011)       (0.011)       (0.011)
Lagged Tobin's Q                0.019 ***     0.019 ***     0.019 ***
                               (0.002)       (0.002)       (0.002)
Nonpension Cash Flow            0.092 ***     0.092 ***     0.092 ***
                               (0.007)       (0.007)       (0.007)
Intercept                       0.016 ***     0.016 ***     0.016 ***
                               (0.003)       (0.003)       (0.003)
Firm fixed effects               Yes           Yes           Yes
Number of observations         11,013        11,013        11,013
Adjusted [R.sup.2]              0.14          0.14          0.14

                                 (4)           (5)          (6)

PDA1

PDA1 x MC

PDA2

PDA2 x MC

ABN_PROD

ABN_PROD x MC

ABN_DISX                      -0.001
                              (0.001)
ABN_DISX x MC                  0.103
                              (0.088)
ABN CFO                                      0.001 ***
                                            (0.000)
ABN_CFO x MC                                -0.335 **
                                            (0.141)
RM                                                        0.001
                                                         (0.001)
RM x MC                                                   0.011
                                                         (0.106)
MC                            -0.530 ***    -0.241 *     -0.530 ***
                              (0.143)       (0.128)      (0.143)
Lagged Funded Status           0.073 ***     0.068 ***    0.073 ***
                              (0.013)       (0.011)      (0.013)
Lagged Tobin's Q               0.015 ***     0.020 ***    0.015 ***
                              (0.002)       (0.002)      (0.002)
Nonpension Cash Flow           0.111 ***     0.093 ***    0.111 ***
                              (0.012)       (0.007)      (0.012)
Intercept                      0.019 ***     0.016 ***    0.019 ***
                              (0.004)       (0.003)      (0.004)
Firm fixed effects              Yes           Yes          Yes
Number of observations         4,463        11,013        4,463
Adjusted [R.sup.2]             0.15          0.14         0.15

Panel B. Earnings Management and Corporate Investment--GMM Estimation

                                 (1)            (2)           (3)

PDA1                           -1.631 ***
                               (0.347)
PDA1 x MC                       9.132
                             (113.550)
PDA2                                         -1.545 ***
                                             (0.307)
PDA2 x MC                                    29.201
                                            (96.669)
ABN_PROD                                                   -0.050 ***
                                                           (0.016)
ABN_PROD x MC                                               2.282
                                                           (1.642)
ABN_DISX

ABN_DISX x MC

ABN_CFO

ABN_CFO x MC

RM

RM x MC

MC                             -1.207 *      -1.044        -0.283
                               (0.714)       (0.732)       (0.209)
Lagged Funded Status           -0.076 *      -0.071 *       0.064 ***
                               (0.043)       (0.040)       (0.014)
Lagged Tobin's Q                0.003         0.003         0.011 ***
                               (0.004)       (0.004)       (0.001)
Nonpension Cash Flow            0.084 **      0.092 ***     0.183 ***
                               (0.038)       (0.034)       (0.015)
Intercept                       0.033 ***     0.031 ***     0.012 **
                               (0.006)       (0.006)       (0.006)
Firm fixed effects               Yes           Yes           Yes
Number of observations          9,650         9,650        10,444
[chi square]                  193.98        202.23        971.91
p-Value                         0.00          0.00          0.00

                                 (4)           (5)          (6)

PDA1

PDA1 x MC

PDA2

PDA2 x MC

ABN_PROD

ABN_PROD x MC

ABN_DISX                      -0.410 *
                              (0.230)
ABN_DISX x MC                 72.781
                             (46.987)
ABN_CFO                                     -0.671
                                            (0.790)
ABN_CFO x MC                                -8.159 *
                                            (4.508)
RM                                                       -0.002
                                                         (0.007)
RM x MC                                                  -2.289
                                                         (6.329)
MC                            -2.693 **      1.105       -0.561 *
                              (1.200)       (1.886)      (0.325)
Lagged Funded Status           0.049         0.004        0.026 **
                              (0.030)       (0.072)      (0.013)
Lagged Tobin's Q               0.012 *      -0.004        0.001
                              (0.007)       (0.023)      (0.001)
Nonpension Cash Flow           0.135 ***     0.160 ***    0.187 ***
                              (0.039)       (0.062)      (0.010)
Intercept                      0.024 ***     0.037 **     0.036 ***
                              (0.008)       (0.017)      (0.002)
Firm fixed effects              Yes           Yes          Yes
Number of observations         3,971        10,444        3,971
[chi square]                  63.93        469.81        20.13
p-Value                        0.00          0.00         0.01

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table IX. The Relation between Earnings Management and DB Plan
Contributions

The table reports the results of univariate analysis of the relation
between DB plan contributions scaled by the book value of assets and
noncash-generating earnings management (NEM) and cash-generating
earnings management (CEM) for subgroups of firms based on
independent double sorts. The regression model is Equation (10). The
dependent variable, corporate investment, is measured as capital
expenditures divided by book assets at the beginning of the year.
Measures of NEM and CEM are described in the text and Appendix A.
The independent variables PDA1, PDA2, and ABN_PROD are alternative
measures of NEM. CEM measures include ABN_DISX, ABN_CFO, and RM. DB
sponsoring firms with positive MC in the previous year are first
sorted into High MC and Low MC subgroups based on their MC scaled by
the book value of assets relative to the sample median. Then, these
firms are sorted into the High EM and Low EM subgroups based on
their NEM or CEM measures, relative to the respective sample
medians. DB plan contributions are extracted from Form 5500 that DB
plan sponsors file with the Internal Revenue Service (IRS) annually
and aggregated to firm level. The t-tests are used to make
statistical inference about the differences in the means of DB plan
contributions between the High EM and Low EM subgroups, both belong
to the High MC subgroup.

                     PDA1               PDA2             ABN_PROD

Subgroups         N       Mean       N       Mean       N       Mean

Panel A. DB Contributions--High MC

Low EM          1,383     0.016    1,381     0.016    1,399     0.016
High EM         1,366     0.017    1,368     0.017    1,348     0.016
Difference               -0.001             -0.001              0.000
t-Statistics             -0.952             -0.714              0.431
p-Value                   0.34               0.48               0.67

Panel B. DB Contributions--Low MC

Low EM          1,366     0.004    1,368     0.004    1,348     0.004
High EM         1,384     0.004    1,382     0.004    1,399     0.004
Difference                0.000              0.000              0.000
t-Statistics              0.811              1.147              0.901
p-Value                   0.42               0.25               0.37

                   ABN_DISX           ABN_CFO            ABN_RM

Subgroups         N       Mean       N       Mean       N       Mean

Panel A. DB Contributions--High MC

Low EM           506      0.015    1,328     0.015     501      0.015
High EM          493      0.017    1,421     0.017     498      0.017
Difference               -0.002             -0.002             -0.002
t-Statistics             -1.729             -2.030             -1.668
p-Value                   0.08               0.04               0.10

Panel B. DB Contributions--Low MC

Low EM           503      0.004    1,421     0.004     508      0.004
High EM          516      0.004    1,329     0.004     511      0.004
Difference               -0.000             -0.000             -0.000
t-Statistics             -0.808             -0.492             -0.139
p-Value                   0.42               0.62               0.89
COPYRIGHT 2017 Financial Management Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Phan, Hieu V.; Khieu, Hinh D.; Golec, Joseph
Publication:Financial Management
Article Type:Report
Date:Mar 22, 2017
Words:21115
Previous Article:Growing pains: international instability and equity market returns.
Next Article:Labor rights, venture capital, and firm performance.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters