Federal Government Contracts and Financial Reporting Quality

Chris He; Mark J. Kohlbeck

doi:10.2139/ssrn.3807526

Federal Government Contracts and Financial Reporting Quality

He, Chris;Kohlbeck, Mark J. 2021-03-18 00:00:00 Government spending is essential for the US economy, and the amount of capital that flows from the government to US firms increased substantially in recent years. Despite the economic importance of the corporate-government contracting relationship, we know little about the firm- level financial outcomes associated with government contracts. Federal government regulations require federal contractors to maintain strong internal control over financial reporting and government contractors have strong incentives to maintain government contracts. As a result, we expect and find that corporate government contracting relationships are associated with higher firm-level financial reporting quality compared to non-government contractors. Further, we find that the improvement of financial reporting quality begins when a firm becomes a federal contractor and greater amounts of government contracts revenue are associated with higher levels of financial reporting quality. We also find the quality of financial reporting weakens after a firm loses government contracts. Collectively, our empirical results suggest that having the government as a customer has a positive impact on the quality of financial reports. Keywords: public procurement, government contractors, financial reporting quality Data Availability: Data are available from the public sources cited in the text. Electronic copy available at: https://ssrn.com/abstract=3807526 Federal Government Contracts and Financial Reporting Quality 1. Introducation US government spending was $4.45 trillion in 2019, representing 21% of the US Gross Domestic Product. This massive amount of spending drives large portions of the US economy and creates millions of jobs for US citizens (Davidson 2018). While government spending is an essential and substantial component of the US economy, we know relatively little about the effect of the government procurement process on the quality of government suppliers’ firm-level financial reporting. It is important to understand federal government contractor’s financial reporting quality since taxpayers are interested in knowing whether public resources are being managed efficiently and effectively. As a result, we investigate whether supplier-customer contracting relationships between corporations and the US Government affect firm-level financial reporting quality. For firms that have contracts with the US government, the contracts account for a large share of the contractors’ revenue. Ellis, Fee, and Thomas (2012) report that 13.4% of public firms' major customers are the US Government and its agencies accounting for one-third of their respective total revenue on average. As government contracts provide contractors a steady stream of revenue that lowers suppliers’ demand uncertainties and increases profitability (Cohen and Li 2019); however, the contractor firms also become increasingly more government- The US federal government awards about $500 billion in contracts annually and is the largest buyer of goods and services in the nation (Samuels 2021). US federal government expenditure average 20% of US GDP over the last 40 years (Cohen and Li 2019). Electronic copy available at: https://ssrn.com/abstract=3807526 dependent (Ellis et al. 2012). Corporate managers therefore have strong incentives to maintain these corporate-government contracting relationships in order to enhance firms’ profitability, improve their stock valuation, and lower their cost of capital (e.g., Cohen, and Li 2019, Banerjee, Dasgupta, and Kim 2008, and Dhaliwal, Judd, Serfling, and Shaikh 2016). Given this incentive, government suppliers provide more voluntary disclosures relative to non-government contractors (Samuels 2021). Existing literature provides evidence as to how major customers affect firm-level financial reporting quality. For example, Raman and Shahrur (2008) provide evidence that firms use earnings management to influence their major commercial suppliers’ or customers’ perceptions of their business prospects. We expect that government contracts impact firms’ reporting outcomes significantly different from commercial contracts (Cohen and Li 2019). We extend this line of literature by investigating whether being a government contractor are associated with the quality of their financial reports provided to the public. We believe that business transactions with the US Government are associated with higher quality financial reporting by government suppliers for a number of reasons. First, the government supplier faces less competition than the typical commercial supplier because of noncompetitive procurement contracts (Mills, Nutter, and Schwab 2013). The unique nature of government sales motivates government suppliers to increase their relationship-specific Government supplier and governement contractor are used interchangably throughout this dissertation to mean a firm that sells goods or services to the government.Further, US Government refers to all departments and agencies of the US Federal Governement. Electronic copy available at: https://ssrn.com/abstract=3807526 investments in order to maximize transaction efficiencies to ensure a competitive position (Cohen and Frazzini 2008) and thereby enhance profitability (Cohen and Li 2019). As a result, corporate managers are less likely to choose an accounting method that involves shifting income from one period to another, resulting in higher financial reporting quality. Second, government agencies have a low default risk and often offer longer-term procurement contracts relative to commercial contracts (Dhaliwal et al. 2016). Low operational and demand uncertainties enhance the forecastability of corporate outcomes and thereby lower corporate managers’ opportunism to manage earnings to meet personal objectives, reducing agency conflicts between managers and shareholders. Third, the government procurement process is administered according to the Federal Acquisition Regulation (FAR) and the government can terminate a contract at any given time (FAR 49.101). Thus, the government agency has greater bargaining and enforcement power to obtain necessary information from its contractors than private-sector customers (Chaney, Faccio, and Parsley 2011). Government contractors will therefore have a strong incentive to ensure that financial reports have the quality needed to mitigate the potential risk of losing government contracts. We regress measures of financial reporting quality on constructs for corporate- government relationships and control variables based on prior research to test our hypothesis. We find a negative association between both the size and existence of government contracts and the level of discretionary accruals as well as the probability of material weakness in internal control, and restatements, supporting our hypothesis. Our study adds to the literature on implications of corporate-government contracting. We extend Samuels (2021) that focuses on whether federal contractors are more likely to issue Electronic copy available at: https://ssrn.com/abstract=3807526 management forecasts and the frequency of the issuance. Our study, on the other hand, provides extensive evidence on an important aspect of financial accounting, that of reporting quality. Also, our study adds to the stream of research about the effects of bilateral relationships on firm internal decisions and performance, especially studies looking at the consequences of major customer dependency (e.g., Kalwani and Narayandas 1995; Banerjee et al. 2008; Gosman and Kohlbeck 2009). Additionally, our study responds to Healy and Wahlen’s (1999) call for future research to further explain how business factors drive accruals by documenting how firms contracting with the federal government and its agencies exhibit a lower level of accruals management. Our findings should be of interest to investors, as they seek to minimize the risk of investing in government contractors' stocks and make strategic decisions regarding their buy and hold decisions. The findings should also be of interest to regulators (e.g., the Securities and Exchange Commission [SEC]) and other stakeholders (e.g., customers and suppliers) who are interested in understanding the factors that can improve firm’s financial reporting quality, as well as stakeholder groups who care about firms’ future growth potential. Finally, our study adds to financial reporting literature by investigating the financial reporting quality of government contractors. Early studies focus on the relationship between government contracting regulation and contractor firm’s accounting outcome (e.g., Pownall 1986; Horwitz and Normolle 1988). Subsequent studies investigate whether government contractors enjoy more subsidies from federal agencies (Callahan, Vendrzyk, and Butler 2012; Cohen and Li 2016); are affected by political sensitivities that influence firm outcomes (e.g., Karpoff, Lee, and Vendrzyk 1999; Mills, Nutter, and Schwab 2013); or shift costs from commercial to government contracts to extract benefits (e.g., Lichtenberg 1992; McGowan and Electronic copy available at: https://ssrn.com/abstract=3807526 Vendrzyk 2002; Chen and Gunny 2014). Our study therefore adds to the literature by investigating the financial reporting quality of government contractors. We organize the study as follows. The next section provides a literature review and hypothesis development. The third section contains research design and variable descriptions. The fourth section provides a discussion of our sample and descriptive statistics. We report the results of tests in the fifth section, which includes a discussion of our sensitivity analyses. The final section provides a conclusion. 2. Literature Review and Hypothesis Development Corporate-Government Contracting Relationships and Firm Outcomes Early studies investigating the profitability of government contractors find that these contractors earned higher profits from government contracts. However, these studies provide mixed evidence as to why government contractors are able to generate abnormally high profits. For instance, Reichelstein's (1992) findings suggest that government contractors earn an abnormally high profit from their government business by shifting and charging overhead and pension costs from commercial operations to their government contracts. Lichtenberg (1992) confirms these findings and further suggests that government contractors earn almost three times more profit than commercial contractors. His results indicate that government contractors are also significantly less capital-intensive than non-government contractors. However, McGowan and Vendrzyk (2002) find no evidence that government contractors’ abnormal profits are attributable to cost-shifting. Their study finds that abnormal profits result from price renegotiation with the government agency or cost reimbursement. McGowan and Vendrzyk (2002) argue that the unusually high profitability of government contractors is more likely attributable to non-accounting explanations, such as the lack of industry competition. Electronic copy available at: https://ssrn.com/abstract=3807526 Recent literature investigates factors that can influence a government agency’s decisions to grant contracts to firms. Flammer (2018) investigates the association between corporate social responsibility (CSR) and the likelihood of gaining government procurement contracts. The author argues that a prospective contractor uses CSR reporting to signal it’s quality and differentiate itself from other competitors in order to influence a government agency’s purchasing decision. His results indicate that CSR firms receive more government contracts when the contractor is in a competitive industry. However, firms that make larger political contributions are more likely to receive government contracts with favorable terms (Ferris, Houston, and Javakhadze 2019). Other research investigates the effect of the corporate-government contracting relationships on corporate outcomes. This research generally finds that firms benefit from having a government agency as their major customer because investors and creditors view these firms as having a lower risk than firms without government contractors. Government contractors hold significantly less cash and have less volatile future earnings than non-governmental contractors (Cohen and Li 2016). Government contractors also enjoy a lower number of debt covenants and are less likely to have performance pricing provisions in their loan contracts (Cohen, Li, Li, and Lou 2016). More importantly, government contractors tend to have lower costs of equity capital (Dhaliwal et al. 2016). Cohen and Li (2019) find that firm profitability increases with a concentration of major government customers but decreases with a concentration of major corporate customers. The authors conclude that major government customers have unique and significant influences on firm outcome which differs from major commercial customers. Other research suggest that the government plays a useful role in the contract process by monitoring its contractors. Samuels (2021) investigates whether government monitoring of the Electronic copy available at: https://ssrn.com/abstract=3807526 contractor’s internal information process strengthens the firm’s external reporting quality. Examining the frequency of voluntary disclosures, speed of earnings announcements, and quality of public information about the contractor, Samuels suggests that having both a government contract and a high dollar government contract are associated with the contractors having higher quality of external reporting. Moreover, she finds that external reporting quality improves when a firm is granted a government contract compared to firms that are not awarded government contracts. Cheng et al. (2019) find that government contractors issue more precise and accurate management earnings forecasts than non-government contractors. Their results suggest that government agencies have greater bargaining and enforcement power, which enables them to obtain more information about the supplier firm. Hypothesis Development Government contractors are required to submit annual audited financial statements and compliance reports, including a single audit report (CFR 200.331). FAR requrie federal contractors maintain internal control systems that are suitable for the size of the organization, able to timely discover and disclose control deficiencies, and ensure correction actions are carried out (FAR 3.1). Furthermore, the contractor’s independent auditor has to identify whether there is a material weakness in internal control, any instance of material non-compliance with the Federal Code, any questionable costs related to the federal contract and/or, large and unusual amounts of federal contract-related expenses in comparison to total operating costs. As a result of the additional requirements stipulated by the regulations, the corporate-government contracting relationship is likely to be associated with higher financial reporting quality. Major customers, including the federal government, have the ability and incentive to monitor their key suppliers in order to mitigate potential agency costs (Itzkowitz 2013, and Electronic copy available at: https://ssrn.com/abstract=3807526 Itzkowitz 2015). FAR regulations extensively formalize government monitoring policies and procedures for government agencies. These policies and procedures enhance government agency’s abilities to monitor contractors’ internal control and information processing. As a result, effective government monitoring deters earnings management as the government can terminate a contract at any given time if it identifies corporate wrongdoing (FAR 49.101). Government agencies also constantly update their policies and procedures to ensure the monitoring is adequate. Further, according to the positive accounting theory, firms will choose accounting policies that minimize contract costs (Watts and Zimmerman’s 1978). The political cost hypothesis therefore predicts that firms subject to government scrutiny such as government contractors take actions to minimize the probability of negative government actions. Corporate- government contracting is expected to be associated with higher firm-level financial reporting quality. Competition for government contracts is generally less than typical commercial transactions (Mills, Nutter, and Schwab 2013). Government contractors are therefore motivated to make relationship-specific investments to maximize transaction efficiencies. As a result, government contractors enjoy a stronger competitive position (Cohen and Frazzini 2008) and experience enhanced profitability (Cohen and Li 2019). Therefore, there are lower incentives to choose accounting methods that shift income or otherwise lower financial reporting quality. According to CFR 200.331, government contractors are requried to disclose their financial information in order to maintain their government contracts. Electronic copy available at: https://ssrn.com/abstract=3807526 Consistent with this argument, recent studies show that government agencies positively impact firm external reporting as management earnings forecasts are more precise and accurate (Cheng et al. 2019), information asymmetry is reduced, and the firm overall transparency is enhanced (Samuels 2021). Government contractors also benefit from the predictability of operating cash flows from government contracts, the lower default risk, and long-term nature of the contracts. Dechow and Dichev (2002) show that a higher certainty of future cash flows is associated with a higher accrual quality. Accrual quality is also positively associated with earnings persistence. Therefore, government contractors are less likely to smooth earnings in order to signal private information about future performance to stakeholders (Watts and Zimmerman 1986, Sankar and Subramanyam 2001; Tucker and Zarowin 2006). Based on the above arguments, our hypothesis predicts that corporate-government contracting relationships enhance firm-level financial reporting quality. Our hypothesis is stated as follows: HYPOTHESIS: Corporate-government relationships are associated with higher firm-level financial reporting quality. However, a negative association is possible given that firms dealing with government agencies are incentivized to disclose optimistic financial reports. Agency theory suggests that self-interested contracting parties have the incentive to behave opportunistically in situations of uncertainty and information asymmetry (e.g., Jensen and Meckling 1976 and Jensen 1986). Consistent with the earnings management literature (e.g., Raman and Shahrur 2008), a government contractor might use accruals to inflate reported earnings in order to signal to government agencies that it is able to continue to meet its contractual obligations. Electronic copy available at: https://ssrn.com/abstract=3807526 3. Research Design Corporate-Government Contracting Variables We use the Federal Procurement Data System (FPDS) for data on government contractors. The FPDS contains information on all government contracts and contract modifications beginning in 2000. The US government discloses detailed information on government contracts that are above $25,000, including the total contract value, information about the contractor, and the federal agency. Using this data, we identify firms that have a corporate-government contracting relationship in a given year (GovContract). We also use ContractSize, a continuous measure of contract award size relative to a firm’s total sales (e.g., Mills, Nutter and Schwab 2013; Samuels 2021). The impact of the corporate government contracting relationship on the contractor’s financial reporting quality might vary with the importance of the government contracts to the contractor. Because ContractSize is heavily right-skewed, we follow Samuels (2021) and transform the variable into decile ranks scaled from 0 to 1. This transformation has the advantage of being robust to both outliers and nonlinearities and eases the interpretation of the results. The Federal Procurement Data System data is available at USAspending.gov. The estimated regression coefficient measures the change in the respective measure of financial reporting quality when moving from the bottom decile to the top decile of contract size, ceteris paribus). Electronic copy available at: https://ssrn.com/abstract=3807526 Financial Reporting Quality Variables The first construct for financial reporting quality is discretionary accruals. We follow Kothari, Leone, and Wasley’s (2005) approach for performance-matched discretionary accruals measure. This performance-matched discretionary accrual model adjusts for industry and year performance. We first construct a matched sample based on the firm’s closest return on assets (ROA) in the same year and then estimate the following model annually in each three-digit SIC industry with more than 15 observations. TA , /A = β (1/A𝑖 𝑡 ) + β (∆REV𝑖 𝑡 /A 𝑖 𝑡 − ∆REC𝑖 𝑡 /A 𝑖 𝑡 ) + β (PPE𝑖 𝑡 /At−1) + i t i,t-1 1 , −1 2 , , −1 , , −1 3 , β (ROA ) + β (MB ) + 𝜀 (1) 4 𝑖 ,𝑡 5 𝑖 ,t 𝑖 ,t where variables are defined as follows: TA is computed as earnings before extraordinary items minus cash flow from operating activities from the statement of cash flows, A represents total assets, ∆REV is the change in revenues, ∆REC is the change in net receivables, PPE represents the gross amount of property, plant, and equipment, ROA is net income before extraordinary items divided by lagged total assets, and MB is the ratio of the book value of net assets divided by the market capitalization. The residual from Equation (1) proxies for discretionary accruals. Prior studies argue that the modified Jones (1991) model does not capture a firms’ earnings management when financial performance is extreme, resulting in potentially incorrect inferences and model misspecification. Building on the modified Jones model and Dechow et al. (1995), Kothari et al. (2005) use a control sample to create performance-matched discretionary accruals in order to identify abnormal earnings management. Electronic copy available at: https://ssrn.com/abstract=3807526 We then take the absolute value of the discretionary accruals (AbsAccruals) to capture a firm’s financial reporting quality. Greater AbsAccruals indicate poorer financial reporting quality. Our next construct for financial reporting quality is the accruals quality measure developed by Dechow and Dichew (2002) as modified by McNichols (2002). Dechow and Dichev (2002) document that working capital accruals are affected by cash flows by mapping the working capital accruals into operating cashflows. McNichols (2002) modifies the measure by including property, plant and equipment, and changes in sales. Following Ball and Shivakumar (2006) and Dou, Khan, and Zou (2016), we estimate the following model annually in each three- digit SIC industry with more than 15 observations: TA = β + β CFO + β CFO + β CFO + β ΔSALES + β PPE + 𝜀 (2) i,t 0 1 𝑖 ,𝑡 -1 2 𝑖 ,𝑡 3 𝑖 ,𝑡 +1 4 𝑖 ,𝑡 5 𝑖 ,𝑡 𝑖 ,𝑡 where variables are defined as follows: TA is as earnings before extraordinary items minus cash flow from operating activities from the statement of cash flows, CFO is cash flow from operating activities, ΔSALES is the year to year change in total sales, and PPE is the gross amount of property, plant, and equipment. All variables are divided by average total assets. The residual from Equation (2) is the accrual quality measure (AccrualQuality). Greater values of AccrualQuality indicate poorer financial reporting quality. A firm’s financial reporting quality is a reflection of its willingness to put a strong internal control structure in place to ensure that financial reports are of adequate quality. Consequently, our third construct for financial reporting quality is the presence of an internal control material weakness. Internal control material weakness data is collected from the Audit Analytics database. We define ICMW as an indicator variable equal to one if there is a general (systemic) material weakness in internal controls reported under either SOX 302 (management) Electronic copy available at: https://ssrn.com/abstract=3807526 or SOX 404 (auditor) in a given year, and zero otherwise. We also consider the number of material weaknesses reported in our sensitivity tests. Restatements capture the failure of a firm’s financial reporting process. Restatement data is obtained from the Audit Analytics database. Restatement is an indicator variable equal to one if the government contractor’s financial statement for that year is subsequently restated, and zero otherwise. Later sensitivity tests consider intentional misstatements as it is important to distinguish error from irregularities (Hennes, Leone, and Miller 2008). This second restatement measure captures restatements that were followed by fraud-related class action lawsuits. Corporate-Government Relationship and Accruals Earnings Management Our first model tests the relationship between corporate-government contracting and financial reporting quality using discretionary accruals and accruals quality. We follow Ham, Lang, Seybert, and Wang (2017) for our empirical specification modified to include governance variables to explain accruals earnings management (subscripts for firm and year are omitted here and in later equations for brevitiy unless necessary to understand the equation). AbsAccruals (or AccrualQuality) = β + β CORP_GOV + β lnAsset + β M/B + 0 1 2 3 β Leverage + β FirmAge + β Loss + β %ΔCashSales + β ΔROA + 4 5 6 7 8 β CfoVolit + β SalesVol + β E/P+ β lnAnalyst + β CEOChair + 9 10 11 12 13 β ACSize + β Independence + β ACExperts + β Big4 + 14 15 16 17 We exclude the CEO characteristics control variables from Ham et al. (2017). Their paper foucses on CFO and CEO characteristics whereas our paper focuses on firm characteristics. Electronic copy available at: https://ssrn.com/abstract=3807526 β LnAudtenure + Fixed Effects + ɛ (3) where variables are defined in Appendix A. Year and two-digit SIC industry fixed effects are included in our estimation to mitigate the impact of time-invariant year and industry characteristics. Standard errors are clustered by firm. As discussed above, we consider two accruals earnings management variables for financial reporting quality - discretionary accruals (AbsAccruals), and accruals quality (AccrualQuality). The key variable of interest is CORP_GOV that is defined as either the presence or amounts of the corporate-government relationship (GovContract and ContractSize). Consistent with our hypothesis, we expect a negative coefficient on both GovContract and ContractSize. Negative coefficients suggests that a government contractor is less likely to engage in accruals earnings management and hence, have higher financial reporting quality. We include several firm characteristics that might influence the level of accrual earnings management. Larger firms have better financial reporting processes and procedures in place to ensure the quality of accruals (e.g., Kinney and McDaniel 1989; Ge and McVay 2005). We expect the coefficients of lnAsset to be negative, indicating that larger firms are associated with higher accruals quality. Market to book ratio and leverage capture growth opportunities and default risk. High growth firms and firms with high default risk are more likely to choose accounting procedures that shift reported earnings from a future period to the current period (e.g., Watts and Zimmerman 1986), resulting in higher discretionary accruals. However, Jensen (1986) suggests the use of debt increase outside monitoring and therefore reduces managers’ opportunitstic behaviors. We expect the coefficient on M/B to be positive but are unable to predict the coefficient on Leverage. Firm age is an indication of firm maturity where younger Electronic copy available at: https://ssrn.com/abstract=3807526 firms are less likely to have established procedures. Thus, we expect the coefficient of FirmAge to be negative, indicating that older firms have higher accruals quality. Firm financial performance influences the level of accruals management. We control for losses, percentage change in cash sales, and change in return-on-assets We expect firms experiencing financial difficulties may pay less attention to the financial reporting process. Greater accrual management generally accompanies increases in ROA (McNichols 2000 and McNichols 2002). Thus, we expect the coefficients on Loss, %ΔCashSales, and ΔROA to be positive. Dechow and Dichev (2002) suggest that cash flows and sales volatilities are significantly associated with accruals. We expect the coefficients of CfoVolit, and SalesVol to be positive, reflecting management’s incentive to smooth earnings for compensation incentives (e.g., Jayaraman 2008, Garrett, Hoitash and Prawitt. 2014). We also control for the earnings to price ratio in order to reflect investors’ anticipated demand for earnings growth in future periods. However, a higher earnings to price ratio might be associated with higher or lower earnings quality (Ham et al. 2017). Firm governance is also a key determinant of a firm's financial reporting quality. We first control for the number of analysts following the firm in order to capture external monitoring. Higher external monitoring should restrict management’s ability to manage earnings, resulting in a higher level of accruals quality. We expect the coefficient of LnAnalyst to be negative. We then control for CEO duality, where the CEO is the chairman of the board. Agency theory argues that powerful CEOs have managerial discretion to manipulate earnings in order to meet personal goals. We expect the coefficient of CEOChair to be positive. Audit committee financial expertise (ACExpert) and independent boards of directors (Independence) enhance board monitoring to ensure the quality of financial reports (e.g., DeFond, Hann, and Hu 2005; Electronic copy available at: https://ssrn.com/abstract=3807526 Dhaliwal, Naiker, and Navissi 2010; Carcello, Hermanson, Neal, and Riley 2002). Thus, we expect the coefficients on ACExpert and Independence to be negative. Audit committee size may also impact earnings management (Klein 2002); however, the empirical evidence is mixed (e.g., Larcker, Richardson, and Tuna 2007). As a result, we do not predict the coefficient of ACSize. We control for Big 4 auditors since prior studies find Big 4 auditors have a significantly positive impact on financial reporting quality (e.g., Carcello et al. 2011). Finally, we include auditor tenure (LnAudtenure); however, we are unable to predict the sign of the coefficient since the impact of auditor tenure on audit quality is uncertain (Singer and Zhang 2018). Corporate-Government Relationship and Internal Control Material Weakness Consistent with FAR requirements, we expect government contractors to have stronger internal control structures that improve the quality of financial reports. For our internal control material weakness model, we start with Ham et al. (2017) and add controls for governance. Our specification is as follows: Prob ( ICMW = 1) = F [ δ + δ CORP_GOV + δ lnAsset + δ M/B +δ Leverage + 0 1 2 3 4 δ FirmAge + δ %Loss + δ SalesGrowth + δ Inventories + δ LnAuditFee + 5 6 7 8 9 δ LnAnalyst + δ CEOChair + δ ACSize + δ Independence + 10 11 12 13 δ ACExperts + δ Big4 + δ LnAudtenure + Fixed effects + ɛ ] (4) 14 15 16 where variables are defined in Appendix A and F is a logistic function. We again include year, and two-digit SIC industry fixed effects in our estimation to mitigate the impact of time-invariant year and industry characteristics. Consistent with our hypothesis, negative coefficients on the two CORP_GOV constructs suggest that government contracts are less likely to receive an ICMW, corresponding to a higher level of financial reporting quality. Electronic copy available at: https://ssrn.com/abstract=3807526 Similar to Equation (3), we include firm size, market-to-book ratio, leverage, and firm age. Prior research indicates that larger, growth, older, and firms with lower leverage are associated with higher financial reporting quality (e.g., Ge and McVay 2005, Watts and Zimmerman 1986), Positive (negative) coefficients are therefore expected for Leverage (LnAssets, M/B and FirmAge). We follow prior studies (e.g., Ham et al. 2017; Krishnan 2005 and Ashbaugh‐Skaife, Collins, and Kinney 2007) and control for firm operating risk and business complexity using the number of years that the firm reports negative earnings over the prior four years, sales growth, and inventory. A risky firm with higher business complexity is associated with a higher probability of internal control weaknesses. Thus, we expect %Loss, SalesGrowth, and Inventory to be positively associated with ICMW. We also use audit fees to control for auditor effort; however, we are unable to predict the sign of audit fee since there is a large array of risk factors that are priced into audit fees and prior research documents mixed results between audit fees and audit quality (DeFond and Zhang 2014). Finally, we include a number of variable to control for firm governance. Stronger governance is associated with a lower probability of internal control material weaknesses. Accordingly, we expect inverse relationships for the number of analysts following the firm, audit committee expertise, board of director independence, and Big 4 auditor and a positive association for CEO – chairman duality. Consistent with prior research, we leave the prediction unsigned for audit committee size and auditor tenure. Corporate Government Relationship and Restatements Our final construct for the quality of financial statements is Restatements. We again follow Ham et al. (2017) for our restatement model. To the extent that firms with government Electronic copy available at: https://ssrn.com/abstract=3807526 contracts have a lower tendency to misreport, we expect to find a decreased likelihood that financial statements are misreported. Firms rarely restate their financial statements, and restatements do not capture all types of misreporting. However, the restatement tests represent a more objective outcome of misreporting. We use the following Logit model for our test: Prob (Restatement =1) = F [δ + δ CORP_GOV + δ lnAsset + δ M/B + δ Leverage + 0 1 2 3 4 δ FirmAge + δ ΔInventories + δ ΔReceivables + δ LnAuditFee + 5 6 7 8 δ %ΔCashSales+ δ E/P δ ΔROA + δ LnAnalyst + δ CEOChair+ 9 10 + 11 12 13 δ ACSize + δ Independence + δ ACExperts + δ Big4 + 14 15 16 17 δ LnAudtenure + Fixed effects + ɛ ] (5) where variables are defined in Appendix A. We include year and two-digit SIC industry fixed effects to mitigate the impact of time-invariant year and industry characteristics. Consistent with our hypothesis, a negative coefficient on our two constructs for CORP_GOV suggests that a government contractor is less likely to restate the financial statements and hence, have higher financial reporting quality. We again include firm size, market-to-book ratio, leverage, and firm age in this financial reporting quality specification. We expect that larger, growth, older, and less levered firms are associated with higher financial reporting quality and therefore a lower probability of restatements. Prior research suggests that an increase in working capital and a higher percentage changes in cash sales is associated with earnings management and a higher probability of future restatements (e.g., Beneish 1999; Dechow, Ge, Larson and Sloan 2011). We therefore include ΔInventories, ΔReceivables, and %ΔCashSales. Audit fees are associated with a greater likelihood of restatements consistent with increased audit risk (e.g, Armstrong, Blouin, Electronic copy available at: https://ssrn.com/abstract=3807526 Jagolinzer and Larcker 2015). We also expect higher earnings to price ratio and lower changes in ROA are associated with restatements (e.g., Dechow et al. 2011). Finally, the governance variables included in the discretionary accruals models are included. As the number of analysts following the firm, the proportion of financial experts on the audit committee increases, or the firm has a Big 4 auditor, we expect the probability of restatement declines. However, we make no prediction for the remaining governance variables. 4. Sample Selection and Descriptive Statistics Sample Selection We start with the cross-section of the CRSP and Compustat datasets to obtain our required financial and accounting data over the period from 2001 to 2017. Following prior studies, we exclude firms in highly regulated industries (banking and utilities), which results in 115,572 firm-year observations. We then combine these observations with our government contractor data we assembled using the FPDS. As the contract data does not contain common identifiers such as GVKEY or PERMNO, we perform a fuzzy-match of the parent company names to the names of Compustat companies, and hand-check identified pairs to ensure proper matching. We aggregate all government contract information by matched firm and year. We eliminate 41,392 and 8,712 firm-year observations with missing data to compute the discretionary accruals and control variables, respectively. Next, we eliminate an additional 6,496 firm-year observations with data missing from Audit Analytics. This provides a sample size of 58,972 firm-year observations to estimate the discretionary accruals and restatement models. Since the internal control material weakness data is only available beginning in 2004, we use the second sample of 47,276 firm-year observations to estimate the internal control material weaknesses models. Table 1 presents the sample construction. Electronic copy available at: https://ssrn.com/abstract=3807526 <Insert Table 1> Table 2, Panel A provides the sample distribution by year. The annual distribution is fairly constant across years between the government contractor and total samples. The total government contracts per year in our sample range from $75 billion in 2001 to $200 billion in 2008. Consistent with prior studies (e.g., Mills, Nutter, and Schwab 2013, and Samuels 2021), the annual government contract value averages about 2 to 4 percentage of total firm sales. <Insert Table 2> Sample distributions by the Fama-French 12 industry classifications are presented in Panel B of Table 2. Our sample covers a wide range of industries with computer and software firms representing the largest industry. The computer and software firms also are the most represented among the government contractor firms. Government contractors represent between 21% and 35% of the number of firms for most industries. In terms of government contract value, manufacturing and computers and software are the largest industries ($1.1 trillion, and $700 billion, respectively). The relative importance of government contracts (as a percentage of total sales) is also the greatest in these two industries at 9.0% and 4.6%, respectively. Descriptive Statistics Panel A of Table 3 presents descriptive statistics for the model variables. The mean value of CorpGov indicates that about 27% of the sample firms are government contractors. On average, the sample firms are characterized as large and profitable indicated by mean total assets of approximately $249 million and mean change of 0.5% in return on assets. The sample firms report total debt averaging 23.8% of total assets and lower inherent risk as inventory represents only 11% of total assets. The average firm is just over ten years old. The mean audit committee size and board independence for the sample firms are 3.9 directors and 66.7%, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Furthermore, 29% of audit committee members have financial expertise, and 67.7% of our sample firms are audited by one of the Big 4 auditors. Overall, our sample statistics are similar to those reported by prior research (e.g., Mills, Nutter, and Schwab 2013, and Samuels 2021). In Panel B of Table 3, we compare firm-year observations with and without government contracts. The univariate tests show that firms with government contracts are characterized by significantly lower discretionary accruals and a lower probability of restatements than non- government suppliers. This significant differences in discretionary accruals and restatement numbers provide initial support for our hypothesis. Government contractors are also significantly larger, more profitable, more mature, stronger corporate governance, and subject to more external monitoring. <Insert Table 3> The Pearson correlation coefficients for the model variables are presented in Table 4. The correlations between discretionary accruals and government contracting variables and the correlation between restatements and government contracting variables are negative and significant at a 1% level, consistent with our hypothesis. All significant pair-wise correlations are less than 0.50, with the exception of correlations among three size related variables (lnAsset, lnAnalyst, and lnAuditfee) and among three audit committee variables (ACSize, ACExperts, and Independence), all of which are expected. We also conduct multicollinearity tests and find that VIFs for the independent variables are well below 5; multicollinearity should not be a concern in our setting. <Insert Table 4> Electronic copy available at: https://ssrn.com/abstract=3807526 5. Empirical Results Accruals Management Table 5 report the results of estimating Equation (3) using ordinary least squares regression (OLS). The reported t-statistics are based on robust standard errors adjusting for clustering at the firm level to address heteroscedasticity. Panel A of Table 6 reports results where AbsAccurals is the dependent variable and the government contracting test variable is either GovContract (Model 1) or ContractSize (Model 2). The adjusted R for the estimation is approximately 18% for both models. The estimated coefficient on GovContract is negative and significant (-0.009, p-value < 0.01) in Model 1, which is consistent government contractors being associated with a lower level of the absolute value of discretionary accruals. The estimated coefficient on the ContractSize is negative and significant (-0.011, p-value < 0.01), consistent with the argument that the government contract values are associated with higher financial reporting quality. These results suggest that the absolute level of discretionary accruals for government contractors are 1% lower than non-government contractors. These results are economical significant given the average total asset among government contractors are $540 millions and a one percent decrease is equivalent to $5.4 million less discretionary accruals. Our hypothesis is supported. For the control variables, we observe negative and significant coefficients on lnAsset, FirmAge LnAnalyst, Independence, ACExperts, and Big4 consistent with our expectation that Average total asset for government contracotrs is 6.293 from Table 3 panel B. Exp(6.239) = 540. Electronic copy available at: https://ssrn.com/abstract=3807526 older, larger firms and those with stronger governance are associated with lower discretionary accruals. However, we observe a negative and significant coefficients on SalesVol and E/P. The coefficient on SalesVol is only marginally significant but provides some indications that higher sales volatility is associated with lower discretionary accrual during our sample period. Although we did not predict the sign for E/P, this finding is consistent with Ham et al. (2017). Moreover, the coefficients on Leverage, ACSize, and LnAudtenure are positive and significant, suggesting that high leverage firms and those with larger audit committees and longer auditor tenure are associated with higher discretionary accruals. We also observe positive and significant coefficients on, M/B, Loss, %∆CashSales, CfoVolit, and CEOChair in line with our prediction that growth firms, firms with operational issues or weaker governance are associated with higher discretionary accruals. Panel B of Table 5 reports results using AccrualQuality as the dependent variable where we vary the test variable between GovContract (Model 1) and ContractSize (Model 2). The adjusted R for the both estimations is approximately 3%. The estimated coefficient on GovContract is negative and significant (-0.002 in Model 1, p-value < 0.01), and the estimated coefficient on ContractSize is negative and significant (-0.003 in Model 2, p-value < 0.01). These results are,consistent with our hypothesis that government contractors are associated with higher accruals quality. As for our control variables, larger firms are associated with higher accruals quality. Firms that are experiencing operational difficulties and higher cashflow volatility are associated with lower accrual quality. Overall, our results are consistent with those reported by Ham et al. (2017). <Insert Table 5> Electronic copy available at: https://ssrn.com/abstract=3807526 Internal Control Weaknesses Table 6 present the results of estimating Equation (4) using Logit. The reported z- statistics are based on heteroscedasticity robust standard errors adjusting for clustering at the firm level and include industry and year fixed effects. Pseudo R for the estimations are approximately 11% for both models. The estimated coefficients on GovContract and ContractSize are negative and significant (-0.260, p-value < 0.01;-0.309, p-value < 0.01, respectively). Our results support our hypothesis and are consistent with the argument that both government contractors and the contract value are associated with a lower probability of internal control weaknesses. Firms that are larger, more mature, profitable, higher leveraged, with more AC experts, or audited by Big 4 firm are associated with lower probability of internal control material weakness. Also, higher audit fees and longer auditor tenures are associated with higher probability of internal control material weakness. <Insert Table 6> We estimate a multinomial logit regression using the number of internal control material weaknesses s as an alternative dependent variable (not tabulated). We find results consistent with those reported in Table 6. The estimated coefficients on both GovContract and ContractSize are negative and significant (-0.263, p-value < 0.01; -0.313, p-value < 0.01, respectively), The sample size is reduced to 47,185 firm-year observations as the internal control weakness data coverage start at 2004.. Electronic copy available at: https://ssrn.com/abstract=3807526 consistent with the argument that the government contractors are associated with a lower number of internal control weaknesses, supporting our hypothesis. Restatements Table 7 presents the set of tests for financial reporting quality by estimating Equation (5) for Restatements. The reported z-statistics are based on robust standard errors at the firm level to address heteroscedasticity. The pseudo R for the estimation is approximately 8% for both models. The estimated coefficients on the GovContract and ContractSize are both negative and significant (-0.199, p-value < 0.01; -0.235, p-value < 0.01, respectively). These results are consistent with the government contractors being associated with a lower probability of restatements, supporting our hypothesis. Consistent with findings on internal control material weaknesses, we find the probability of restatement is lower for firms that are larger, more mature, have higher earnings to price ratios, more analysts following, more AC experts, or are audited by a Bif 4 firm. Also, greater leverage, audit fees and percentage change in cash sales are associated with a higher probability of restatements. <Insert Table 7> Intentional mistatements are considered as another proxy for financial reporting quality. Our dependent variable in this model takes the value of one if the government contractor’s financial statements are subsequently restated because of a fraud-related class action lawsuit, and zero otherwise. Our results (untabulated) for the estimated coefficients on the GovContract and ContractSize are both negative and significant (-0.338, p-value < 0.05; -0.420 p-value < 0.05, respectively). These results are consistent with those reported in Table 7 for Restatements and Electronic copy available at: https://ssrn.com/abstract=3807526 the argument that government contractors are associated with lower probabilities of financial statement fraud, supporting our hypothesis. Additional Analyses In the following paragraphs, we discuss a number of additional analysis to further validate the association of government contracts with firm-level financial reporting quality. Change Analysis and Government Contractors We perform a analysis on the effects of change in government contract size on the change in discretionary accruals. Table 6 suggests that government contract size exhibits downward pressure on the contractor's discretionary accruals, suggesting that the association with higher firm-level financial reporting quality may be dependent on contract size. We therefore limit our sample to government contractors and use the change in government contract sales as a percentage of the contractor's total firm sales as our variable of interest. The advantage is that this approach provides more reliable and robust evidence on the effect of government contracts on firm-level financial reporting quality. We report the estimation results in Table 8 and find consistent results across the two financial reporting measures, a significant negative relationship between the change of government contract size and the change in discretionary accruals. The estimated coefficients on the ∆ContractSize are negative and significant (-0.018, p-value < 0.05; -0.009, p-value < 0.05), consistent with our hypothesis. As contract size increases, discretionary accruals are lower and accounting quality is higher. <Insert Table 8> Propensity-Score Matching Model Electronic copy available at: https://ssrn.com/abstract=3807526 In order to reduce concerns regarding misspecification of the functional form whereby the treatment of government contractors is dissimilar to the treatment of non-government contractors, we use propensity-score matching (PSM) to provide stronger controls for the effects of various firm characteristics that are separate from our variables of interest. Although Shipman, Swanquist and White (2017) advise that this method does have weaknesses, we use PSM in combination with our main model to control for the effects of observable firm characteristics and increase the validity of our primary findings. Following the PSM model suggested by Shipman et al. (2017), we estimate a Logit model explaining the probability of being a government contractor using the control variables from Equation (3). Using a caliper distance matching method with a maximum distance of one percent and without replacement, we match firms that were government contractors in any given year with firms that were never government contractors during our sample period. This procedure produces a sub-sample of firms which do not differ based on pre-existing firm characteristics but have a different treatment effect (i.e., GovContract = 1 vs. GovContract = 0). Therefore, the difference in the two subsamples is whether a firm is a government contractor or not, and provides more robust finding in testing the impact of contracting with the government on firm-level financial reporting quality. The matching procedures provide a sample that does not have significant differences in firm characteristics between firms with and without government contracts by comparing the mean and median values of two subsamples (untabulated). Electronic copy available at: https://ssrn.com/abstract=3807526 We estimate Equations (3) to (5) using the PSM sample (untabulated) and our results for the variables of interest (GovContract and ContractSize) are consistent with the results for the full sample across all models, supporting our hypothesis. Overall, these results reduce the aforementioned concerns regarding the misspecification of functional form whereby treatment firms are dissimilar to control firm. Alternative Measures for Financial Reporting Quality We consider three additional financial reporting quality measures to substantiate the robustness of our results. As discussed in the following paragraphs, our results are consistent with our earlier findings the government contracting is associated with higher financial reporting quality. First, prior research recognizes the important role of accounting conservatism for firm’s contracting (e.g., Hui, Klasa, and Yeung 2012). Goh and Li (2011) find that firms with a stronger internal control environment are more likely to understand the role of conservatism in contracting and reducing agency conflicts. Therefore, we expect federal contractors to commit to producing conservative financial statements that reduce the contracting-related transaction costs. We use the C-Score from Khan and Watts (2009) that is based on Basu’s (1997) measure of asymmetric timeliness of earnings. A higher C-Score indicates greater conservtism. We then regress C-Score on the existence of a government contracting relationship and the control variables from Equation (3) and report the results as Model (1) in Table 9. The estimated coefficient on GovContract is positive and significant (0.002, p-value < 0.05) suggesting that government contractors are associated with higher accounting conservtism as compares to non- contractors. <Insert Table 9> Electronic copy available at: https://ssrn.com/abstract=3807526 Second, earnings management can occur through accruals and/or real activities management. The difference is that accruals management misrepresents the firm’s underlying operating performance but does not involve altering firm’s operations. Real activities management involves altering normal business activities with intention to mislead its stakeholders into believing that the reported financial performance is achieved in the normal course of business (Roychowdhury 2006). Since federal government agencies are important stakeholders, investigating whether federal contractors are associated with real activities management is also necessary. We follow Cohen, Dey, and Lys (2008) and combine measures for abnormal discretionary expenses, abnormal cash flows from operations, and abnormal production costs to provide an aggregate real activities earnings management measure (REM) . Higher REM indicates an increasing level of earnings management. We regress REM on the existence of a government contracting relationship and the control variables from Equation (3) and report the results as Model (2) in Table 9. The estimated coefficient on the GovContract is negative and significant (-0.028, p-value < 0.05), suggesting that government contractors are associated with a lower level of real activities earnings management compared to non-contractors. We acknowledge that the three individual real earning management variables have different implications. As result, we also estimate the three individual real earnings management proxies. In untabulated results, we find that federal contractors are significantly associated with lower abnormal discretionary expenses and abnormal cash flow from operations. However, we did not find a significant association between the federal contractor and abnormal production cost. Electronic copy available at: https://ssrn.com/abstract=3807526 Third, Hoitash and Hoitash (2018) find that higher level of financial statement complexity (FSC) is associated with greater probability of restatement risk. Higher FSC is assumed to be attributable to managers’ intentional efforts to obfuscate financial information (e.g., Li, 2008, Lo, Ramos, and Rogo, 2017). We therefore expect government contracting is associated lower FSC. Using Hoitash and Hoitash (2018)’s FSC measure (LnARC), we regress LnARC on the existence of a government contracting relationship and the control variables from Equation (3) and report the results as Model (3) in Table 9. The estimated coefficient on the GovContract is negative and significant (-0.019, p-value < 0.05), suggesting that government contractors are associated with less complex financial statement compared to non-contractors, consistent with lower efforts to obfuscate financial statement information. Endogeneity Concerns A plausible alternative explanation for our results is that federal government agencies only award firms with higher-level financial reporting quality. As a result, high financial reporting quality among government contractors is a pre-existing characteristic of these firms before becoming a government contractor. While we cannot completely rule out this possibility, we conduct a test to mitigate this concern. We argue that government contracts help a firm to improve profitability and forecast ability and therefore, the government contractor is less likely to use discretionary accruals. In this case, we should observe increased discretionary accruals if a The FRC data is publicly available on http://www.xbrlresearch.com/ starting from 2009. Electronic copy available at: https://ssrn.com/abstract=3807526 firm loses government contracts and the opposite to occur if a firm is awarded a government contract. We utilize the PSM sample and estimate a model regressing the year-to-year change in discretionary accruals and accruals quality on indicator variables capturing whether a firm gain or loses government contracts during the previous year. The dependent variable in Model (1) of Table 10 is the change in discretionary accrual (∆AbsAccruals). The estimated coefficient on GovContractLoss is positive and significant (0.020; p-value < 0.01) indicating that when the year after losing government contracts, there is a significant increase in discretionary accruals. On the other hand, the coefficient on GovContractGain is negative and significant (-0.014; p-value < 0.10) indicating when a firm awarded a federal government contract in the prior year, there is a decrease in discretionary accruals. Similar results are obtained when the dependent variable is the change in accruals quality (Model 2). Combined, these results further confirm our previous findings that having a federal government contract is associated with the level of discretionary accruals. [Insert Table 10] In untabulated results, we also consider future weakness in internal control and future restatements. Similar to our results in Table 10, we find a significant increase (decrease) in the likelihood of internal control weakness in the future associated with losses (gains) in government contracts. Results are weaker when the dependent variable is future restatements. We find increased probability of future restatements for gains in government contracts. However, the estimated coefficient on loss in government contracts is negative but insignificant. These results further confirm our previous findings that having a federal government contract influences the probability of future weakness in internal control and restatements. Electronic copy available at: https://ssrn.com/abstract=3807526 Finally, we conduct placebo tests in a manner similar to those conducted by Bertrand and Schoar (2003) in order to address the endogeneity concern that our previous findings are not driven by firms identified as government contractors but instead is driven by a correlated omitted variable. We first exclude government contractor observations and then randomly assign non- government contractors as placebo government contractors in the same percentage as identified government contractors in the original sample. If our previous results are driven primarily by similarities between firms instead of the corporate-government contracting relationship, we should find that the coefficient on the placebo government contractor is also negative and significant. If the association is due to the firm being a government contractor, the placebo government contractor coefficient should not be significant. In untabulated results we repeat our analyses and none of the placebo government contractor coefficients are statistically significant, indicating that the government contractor status drives our results. To further ensure that the procedure is truly random, we repeat this placebo procedure 30 times with similar results. Overall, our placebo tests provide further support for the conclusion that a government contractor plays a role in determining the quality of firms’ financial statements. 6. Conclusion Cohen and Li (2019) and Samuels (2021) suggest that firms with federal government contracts are associated with higher profitability and better information environments. While there is substantial literature relating to firms’ financial reporting quality, the extant literature on investigating the reporting characteristics of government contractors is scarce. Understanding the quality of government contractors’ financial reporting is important because it helps the regulators, investors, and US citizens understand how federal government agencies are allocating Electronic copy available at: https://ssrn.com/abstract=3807526 their resources that contribute to 20% of the US GDP and, on average, 3% of the government contractors total firm sales. Our study investigates different aspects of government contractors’ financial reporting quality, including discretionary accruals, internal control weaknesses, and restatements. In the additional analysis, we further demonstrate government contractors are associated with higher level of accounting conservatism, lower level of real activities earnings management and financial statement complexity. Therefore, we document a comprehensive set of evidence that government contractors are associated with higher quality financial statements than non- government contractors across different measures of financial reporting quality. Our study contributes to the growing body of literature on the characteristics of government contractors. Our results provide important implications for the regulators, investors, and US citizens that government contractors can serve as a defensive investment mechanism that reduces the information risks relative to the non-government contractor by having higher quality financial reports. However, despite our efforts to address any potential endogeneity concerns using the placebo test and PSM matching methodology, we cannot completely rule out the possibility of the corporate government contracting relationships being randomly assigned. Electronic copy available at: https://ssrn.com/abstract=3807526 References Armstrong, C., J. Blouin, A. Jagolinzer, and D. Larcker. 2015. Corporate governance, incentives, and tax avoidance. Journal of Accounting and Economics 60 (1): 1-17. Ashbaugh-Skaife, H., W. Collins, and W. Kinney. 2007. The discovery and reporting of internal control deficiencies prior to SOX-mandated audits. Journal of Accounting and Economics 44 (1-2): 166-192. Ball, R., and L. Shivakumar. 2006. The role of accruals in asymmetrically timely gain and loss recognition. Journal of Accounting Research 44 (2): 207-242. Banerjee, S., S. Dasgupta, and Y. Kim. 2008. Buyer-supplier relationships and the stakeholder theory of capital structure. The Journal of Finance 63 (5): 2507-2552. Basu, S. 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics 24 (1): 3-37. Beneish, M. 1999. The detection of earnings manipulation. Financial Analysts Journal. 55 (5): 24-36. Bertrand, M., and A. Schoar. 2003. Managing with style: The effect of managers on firm policies. The Quarterly Journal of Economics 118 (4): 1169-1208. Callahan, C., V. Vendrzyk, and M. Butler. 2012. The impact of implied facilities cost of money subsidies on capital expenditures and the cost of debt in the defense industry. Journal of Accounting and Public Policy 31 (3): 301-319. Carcello, J., D. Hermanson, T. Neal, and R. Riley. 2002. Board characteristics and audit fees. Contemporary Accounting Research 19 (3): 365-384. Chaney, P., M. Faccio, and D. Parsley. 2011. The quality of accounting information in politically connected firms. Journal of Accounting and Economics 51(1-2): 58-76. Electronic copy available at: https://ssrn.com/abstract=3807526 Chen, H., and K. Gunny. 2014. Profitability and Cost Shifting in Government Procurement Contracts. Available at SSRN: https://ssrn.com/abstract=2482605 Cheng, A. and W. Huang, and S. Zhang. 2019. Major government customer and management earnings forecasts. Working paper, available at SSRN: https://ssrn.com/abstract=3477672. Cohen, D., and B. Li. 2016. Why do firms hold less cash? A customer base explanation. Working paper, available at SSRN: https://ssrn.com/abstract=2440794. Cohen, D., and B. Li. 2019. Customer-base concentration, investment, and profitability: The US government as a major customer. The Accounting Review 95 (1): 101-132. Cohen, D., and B. Li, N Li, and Y. Lou. 2016. Major government customers and loan contract terms. Working paper, available at SSRN: https://ssrn.com/abstract=2868761. Cohen, D., A. Dey, and T. Lys. 2008. Real and Accrual-Based Earnings Management in the Pre- and Post-Sarbanes-Oxley Periods. The Accounting Review 83(3): 757-787. Cohen, L., and A. Frazzini. 2008. Economic links and predictable returns. The Journal of Finance 63 (4): 1977-2011. Davidson, K. 2018. A big reason U.S. econoour is accelerating: Government spending WSJ.online (October 25). Available at: https://www.wsj.com/articles/government-and- military-spending-fuel-u-s-growth-1540459800/. Dechow, P., W. Ge, C. Larson, and R. Sloan. 2011. Predicting material accounting misstatements. Contemporary Accounting Research 28 (1): 17-82. Dechow, P., and We. Dichev. 2002. The quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review 77 (1): 35-59. Electronic copy available at: https://ssrn.com/abstract=3807526 DeFond, M., R. Hann, and X. Hu. 2005. Does the market value financial expertise on audit committees of boards of directors? Journal of Accounting Research 43 (2): 153-193. DeFond, M., and J. Zhang. 2014. A review of archival auditing research. Journal of Accounting and Economics 58 (2-3): 275-326. Dhaliwal, D., S. Judd, M. Serfling, and S. Shaikh. 2016. Customer concentration risk and the cost of equity capital. Journal of Accounting and Economics 61 (1): 23-48. Dhaliwal, D., V. Naiker, and F. Navissi. 2010. The association between accruals quality and the characteristics of accounting experts and mix of expertise on audit committees. Contemporary Accounting Research 27 (3): 787-827. Dou, Y., M. Khan, and Y. Zou. 2016. Labor unemployment insurance and earnings management. Journal of Accounting and Economics 61 (1): 166-184. Ellis, J., E. Fee, and S. Thomas. 2012. Proprietary costs and the disclosure of information about customers. Journal of Accounting Research 50 (3): 685-727. Federal Acquisition Regulation. 2018. 48 C.F.R. § 16. Ferris, S., R. Houston, and D. Javakhadze. 2019. It is a sweetheart of a deal: Political connections and corporate‐federal contracting. Financial Review 54 (1): 57-84. Flammer, C. 2018. Competing for government procurement contracts: The role of corporate social responsibility. Strategic Management Journal 39(5): 1299-1324. Garrett, J., R. Hoitash, and D. Prawitt. 2014. Trust and financial reporting quality. Journal of Accounting Research 52 (5): 1087-1125. Ge, W., and S. McVay. 2005. The disclosure of material weaknesses in internal control after the Sarbanes-Oxley Act. Accounting Horizons 19 (3): 137-158. Electronic copy available at: https://ssrn.com/abstract=3807526 Gosman, M., and M. Kohlbeck. 2009. Effects of the existence and identity of major customers on supplier profitability: is Wal-Mart different? Journal of Management Accounting Research 21 (1): 179-201. Goh, B., and D. Li. 2011. Internal controls and conditional conservatism. The Accounting Review 86 (3): 975-1005. Ham, C., M. Lang, N. Seybert, and S. Wang. 2017. CFO narcissism and financial reporting quality. Journal of Accounting Research 55 (5): 1089-1135. Healy, P., and J. Wahlen. 1999. A review of the earnings management literature and its implications for standard setting. Accounting Horizons 13 (4): 365-383. Hennes, K., A. Leone, and B. Miller. 2008. The importance of distinguishing errors from irregularities in restatement research: The case of restatements and CEO/CFO turnover. The Accounting Review 83 (6): 1487-1519. Hoitash, R., and U. Hoitash. 2018. Measuring accounting reporting complexity with XBRL. The Accounting Review 93 (1): 259-287. Horwitz, B., and D. Normolle. 1988. Federal agency R&D contract awards and the FASB rule for privately-funded R&D. The Accounting Review 63 (3): 414-435. Hui, K., S. Klasa, and E.Yeung. 2012. Corporate suppliers and customers and accounting conservatism. Journal of Accounting and Economics 53 (2): 115-135. Itzkowitz, J. 2013. Customers and cash: How relationships affect suppliers' cash holdings. Journal of Corporate Finance 19: 159-180. Itzkowitz, J. 2015. Buyers as stakeholders: How relationships affect suppliers' financial constraints. Journal of Corporate Finance 31: 54-66. Electronic copy available at: https://ssrn.com/abstract=3807526 Jayaraman, S. 2008. Earnings Volatility, Cash Flow Volatility, and Informed Trading. Journal of Accounting Research 46 (4): 809-851. Jensen, M., and W. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3 (4): 305-360. Jensen, M. C. 1986. Agency costs of free cash flow, corporate finance, and takeovers. The American Economic Review 76 (2): 323-329. Jones, J. 1991. Earnings management during import relief investigations. Journal of Accounting Research 29 (2): 193-228. Kalwani, M., and N. Narayandas. 1995. Long-term manufacturer-supplier relationships: do they pay off for supplier firms? Journal of Marketing 59 (1): 1-16. Karpoff, J., D. Lee, and V. Vendrzyk. 1999. Defense procurement fraud, penalties, and contractor influence. Journal of Political Economy 107 (4): 809-842. Khan, M., and R. Watts. 2009. Estimation and empirical properties of a firm-year measure of accounting conservatism. Journal of Accounting and Economics 48 (2-3): 132-150. Kinney, W, and L. McDaniel. 1989. Characteristics of firms correcting previously reported quarterly earnings. Journal of Accounting and Economics 11 (1): 71-93. Klein, A. 2002. Audit committee, board of director characteristics, and earnings management. Journal of Accounting and Economics 33 (3): 375-400. Kothari, S., A. Leone, and C. Wasley. 2005. Performance matched discretionary accrual measures. Journal of Accounting and Economics 39 (1): 163-197. Krishnan, J. 2005. Audit committee quality and internal control: An empirical analysis. The Accounting Review 80 (2): 649-675. Electronic copy available at: https://ssrn.com/abstract=3807526 Larcker, D., S. Richardson, and We. Tuna. 2007. Corporate governance, accounting outcomes, and organizational performance. The Accounting Review 82 (4): 963-1008. Li, F. 2008. Annual report readability, current earnings, and earnings persistence. Journal of Accounting and Economics 45 (2-3): 221-247. Lichtenberg, F. 1992. A perspective on accounting for defense contracts. The Accounting Review 67 (1): 741-741. Lo, K., F. Ramos. and R. Rogo. 2017. Earnings management and annual report readability. Journal of Accounting and Economics 63 (1): 1-25. McGowan, A., and V. Vendrzyk. 2002. The relation between cost-shifting and segment profitability in the defense-contracting industry. The Accounting Review 77 (4): 949-969. McNichols, M. 2002 Discussion of the quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review 77 (1): 61-69. McNichols, M. 2000. Research design issues in earnings management studies, Journal of Accounting and Public Policy 19 (4-5): 313-345. Mills, L., S. Nutter, and C. Schwab. 2012. The effect of political sensitivity and bargaining power on taxes: Evidence from federal contractors. The Accounting Review 88 (3): 977- Pownall, G. 1986. An empirical analysis of the regulation of the defense contracting industry: The Cost Accounting Standards Board. Journal of Accounting Research 24 (2): 291-315. Raman, K., and H. Shahrur. 2008. Relationship-specific investments and earnings management: Evidence on corporate suppliers and customers. The Accounting Review 83 (4): 1041- Electronic copy available at: https://ssrn.com/abstract=3807526 Reichelstein, S. 1992. Constructing incentive schemes for government contracts: An application of agency theory. The Accounting Review 67 (4): 712-731. Roychowdhury, S. 2006. Earnings management through real activities manipulation. Journal of Accounting and Economics 42 (3): 335-370. Samuels, D. (2021). Government procurement and changes in firm transparency. The Accounting Review 96 (1): 401-430. Sankar, M., and K. Subramanyam. 2001. Reporting discretion and private information communication through earnings. Journal of Accounting Research 39 (2): 365-386. Singer, Z., and J. Zhang. 2018. Auditor tenure and the timeliness of misstatement discovery. The Accounting Review 93 (2): 315-338. Shipman, J.E., Q.T. Swanquist, and R.L. White. 2016. Propensity score matching in accounting research. The Accounting Review 92 (1): 213-244. Tucker, J., and P. Zarowin. 2006. Does income smoothing improve earnings informativeness? The Accounting Review 81 (1): 251-270. Watts, R., and J. Zimmerman. 1978. Towards a positive theory of the determination of accounting standards. The Accounting Review 53 (1): 112-134. Electronic copy available at: https://ssrn.com/abstract=3807526 Appendix A Variable Definitions Corporate-government contracting variables: GovContract Indicator variable equal to one if the company is a government contractor in a given year, and zero otherwise ContractSize Contract award size relative to the firm’s total sales GovContractLost Indicator variable equal to one if the company was a government contractor in the prior year but no longer a government contractor in the current year, and zero otherwise GovContractGain Indicator variable equal to one if the company was a government contractor in the current year but not a government contractor in the prior year, and zero otherwise Financial reporting quality variables: AccrualQuality Standard deviation of residuals from t-4 to t from the Dechow and Dichev (2002) model, as modified by McNichols (2002): CA = α + α CFO + i,t 1 2 i,t-1 α CFO + α CFO + α ΔSALES + α PPE + ɛ where CA is current 3 i, t 4 i, t+1 5 i,t 6 i,t i,t, accruals, CFO is cash flow from operations, ΔSALES is change in sales and PPE is plant, property and equipment AbsAccruals Absolute value of abnormal accruals using the modified Jones (1991) approach and including the firm performance adjustment suggested by Kothari et al. (2005) ICMW Indicator variable equal to one if there is a general (systemic) material weakness in internal control reported under either SOX 302 and SOX 404 in a given year, and zero otherwise Electronic copy available at: https://ssrn.com/abstract=3807526 Restatement Indicator variable equal to one if the government contractor’s financial statement for that year is subsequently restated, and zero otherwise C_Score Conditional conservatism score developed by Khan and Watts (2009) REM Aggregate measure of real activity earnings management that equals the sum of the three standardized real earnings management metrics developed by Roychowdhury (2006) LnARC Natural logarithm of accounting reporting complexity measure developed by Hoitash and Hoitash (2018) Firm-level control variables lnAuditFees Natural logarithm of audit fees lnAsset Natural logarithm of total assets CfoVolit Standard deviation of cash flows divided by total assets over the prior five- year period SalesVol Standard deviation of sales divided by total assets over the prior five-year period E/P Current year earnings before extraordinary items divided by prior year’s market value of common equity M/B Book value of equity divided by the market value of common equity Leverage Ratio of total debt to total assets FirmAge Number of years since the firm first appeared in the CRSP database Loss Indicator variable equal to one if the firm reports losses in a given year, and zero otherwise %ΔCashSales Percent change in sales minus change in accounts receivable in the past year Electronic copy available at: https://ssrn.com/abstract=3807526 ΔROA Change in net income divided by average total assets in the past year LnAnalyst Natural logarithm of the number of analyst following the firm CEOChair Indicator variable equal to one if the CEO also serves as board chair, and zero otherwise ACSize Number of audit committee members ACExperts Proportion of financial experts on the audit committee, where financial expertise is defined as having one of the qualifications in the background of the member: certified public accountant, chief financial officer, chief accounting officer, controller, treasurer or vice-president for finance Independence Proportion of independent members on the board of directors Big4 Indicator variable equal to one if a firm’s auditor is a Big4 auditor, and zero otherwise LnAudtenure Natural logarithm of auditor’s tenure in years %Loss Percentage of loss years over the prior four years SalesGrowth Cumulative percentage change in sales over the prior three years Inventory Inventory divided by total assets ΔInventory Change in inventory divided by average total assets in the past year ΔReceivables Change in accounts receivable divided by average total assets in the past year st th Note: Continuous variables are winsorized at the1 and 99 percentile each year. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 1 Sample Selection Total number of observations in CRSP-Compustat (2001-2017), excluding firms in the banking and utility industries 115,572 Less firm-year observations: Missing data to calculate discretionary accruals 41,392 Missing data for control variables 8,712 Missing Audit Analytics data 6,496 Sample size for discretionary accruals and restatement tests 58,972 Less observations from 2001-2003 11,696 Sample size for internal control material weakness tests 47,276 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 2 Sample Distributions Panel A: Sample Distribution by Year Frequency Frequency of Gov’t Contract Gov’t Contractor % of Gov’t Year of Total % Gov’t % Value Total Sales Sales to Total Sample Contractor ($ millions) ($ millions) Sales 2001 6.28 752 4.71 75,730 1,986,205 3.8% 3,705 2002 6.73 881 5.52 97,929 2,270,550 4.3% 3,967 2003 6.82 959 6.01 114,835 3,133,326 3.7% 4,024 2004 6.72 1,060 6.64 120,039 3,633,572 3.3% 3,963 2005 6.55 1,126 7.06 113,957 4,071,181 2.8% 3,864 2006 6.35 1,138 7.13 170,576 4,623,183 3.7% 3,746 2007 6.19 1143 7.16 159,985 5,273,840 3.0% 3,650 2008 5.85 1108 6.94 199,653 5,355,060 3.7% 3,452 2009 5.63 1049 6.57 161,005 4,696,326 3.4% 3,321 2010 5.49 1018 6.38 148,280 5,070,994 2.9% 3,235 2011 5.31 967 6.06 143,837 5,624,189 2.6% 3,133 2012 5.27 906 5.68 159,579 5,101,447 3.1% 3,105 2013 5.33 847 5.31 109,682 4,922,144 2.2% 3,146 2014 5.58 807 5.06 116,697 4,851,576 2.4% 3,291 2015 5.46 774 4.85 111,875 4,865,324 2.3% 3,218 2016 5.27 775 4.86 129,098 4,571,868 2.8% 3,108 2017 5.16 679 4.25 146,906 4,733,035 3.1% 3,044 Total 58,972 100.00 15,959 100.00 2,279,663 74,783,820 3.0% Electronic copy available at: https://ssrn.com/abstract=3807526 Table 2 (continued) Panel B: Sample Distribution by Industry (Fama-French 12) % of Gov’t Gov’t Frequency Frequency % of Gov’t Contract Contractor Industry of Total of Gov’t Industry Sales to Value Total Sales Sample Contractors Total Total ($ millions) ($ millions) Sales Consumer Non-Durables 3,441 770 22.4% 16,609 5,123,382 0.3% Consumer Durables 1,920 526 27.4% 52,211 4,886,591 1.1% Manufacturing 7,383 2,592 35.1% 1,068,203 11,900,000 9.0% Energy, Oil, and Gas 3,255 214 6.6% 5,315 3,178,877 0.2% Products Chemicals and Allied 2,024 502 24.8% 8,050 2,901,472 0.3% Products Computers and Software 16,608 5,674 34.2% 712,129 15,500,000 4.6% Telephone and TV 9 - 0.0% 16,609 5,123,382 0.3% Transmission Utilities Excluded Wholesale and Retail 6,826 1,440 21.1% 140,743 17,900,000 0.8% Services Healthcare, Medical 10,603 2,627 24.8% 59,778 8,601,541 0.6% Finance Excluded Other 6,903 1,614 23.3% 216,623 4,791,957 4.5% Total 58,972 15,959 27.3% 2,279,663 74,783,820 3.0% Electronic copy available at: https://ssrn.com/abstract=3807526 Table 3 Summary Statistics Panel A: Summary Statistics for Full Sample th th Variable Standard 25 75 (N=58,972) Mean Deviation Percentile Median Percentile AbsAccruals 0.140 0.248 0.031 0.071 0.151 AccrualQuality 0.003 0.093 -0.027 0.000 0.034 IneffControls 0.046 0.209 0.000 0.000 0.000 NumWeaknesses 0.110 0.717 0.000 0.000 0.000 Restatement 0.080 0.271 0.000 0.000 0.000 Fraud 0.014 0.117 0.000 0.000 0.000 C_Score 0.127 0.139 0.054 0.142 0.206 REM -0.156 0.857 -0.443 -0.194 0.032 LnARC 5.203 0.469 4.887 5.223 5.518 CorpGov 0.271 0.444 0.000 0.000 1.000 ContractSize 0.317 0.360 0.100 0.100 0.700 lnAsset 5.520 2.321 3.857 5.522 7.174 M/B 2.514 5.226 1.158 1.631 2.599 FirmAge 10.769 8.715 5.000 9.000 14.000 Leverage 0.238 0.471 0.003 0.147 0.327 Loss 0.426 0.494 0.000 0.000 1.000 CfoVolit 0.084 0.237 0.015 0.035 0.073 SalesVol 0.208 0.479 0.042 0.109 0.219 %∆CashSales 0.183 1.027 -0.063 0.058 0.217 ∆ROA 0.005 0.228 -0.027 0.000 0.023 E/P -0.254 1.563 -0.117 0.019 0.056 CEOChair 0.308 0.461 0.000 0.000 1.000 LnAnalyst 4.516 6.469 0.000 1.778 6.417 ACSize 3.903 1.112 3.000 4.000 4.000 Independence 0.667 0.181 0.571 0.667 0.818 ACExperts 0.293 0.304 0.000 0.250 0.400 Big4 0.674 0.469 0.000 1.000 1.000 LnAudtenure 1.473 0.974 0.693 0.693 2.197 %Loss 0.399 0.384 0.000 0.250 0.750 SalesGrowth 1.988 14.333 -0.061 0.171 0.636 Inventory 0.111 0.131 0.001 0.069 0.170 lnAuditFees 13.221 1.472 12.123 13.252 14.251 ∆Inventory 0.002 0.022 -0.001 0.000 0.007 ∆Receivables 0.004 0.033 -0.005 0.002 0.013 GovContractLoss 0.029 0.168 0.000 0.000 0.000 GovContractGain 0.025 0.156 0.000 0.000 0.000 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 3 (continued) Panel B: Firms with and without Government Contracts Gov’t Contractors Non-Gov’t Contractors p-value (N=15,959) (N=43,013) for Std. Std. Mean t-test Variable Mean Dev. Median Mean Dev. Median Diff. of means AbsAccruals 0.099 0.154 0.058 0.155 0.273 0.078 -0.056 <0.001 AccrualQuality 0.000 0.072 0.000 0.004 0.100 0.000 -0.004 <0.001 IneffControls 0.046 0.209 0.000 0.046 0.209 0.000 0.000 0.963 NumWeaknesses 0.114 0.751 0.000 0.109 0.703 0.000 0.005 0.452 Restatement 0.075 0.263 0.000 0.082 0.274 0.000 -0.007 -0.005 Fraud 0.013 0.112 0.000 0.014 0.118 0.000 -0.001 -0.184 C_Score 0.097 0.14 0.113 0.141 0.137 0.156 -0.044 <0.001 REM -0.262 0.591 -0.248 -0.115 0.938 -0.171 -0.147 <0.001 LnARC 5.249 0.462 5.296 5.184 0.47 5.193 0.065 <0.001 lnAsset 6.293 2.265 6.358 5.233 2.275 5.204 1.060 <0.001 M/B 2.149 1.834 1.630 2.649 6.011 1.632 -0.500 <0.001 FirmAge 12.721 9.764 11.000 10.045 8.174 8.000 2.677 <0.001 Leverage 0.211 0.366 0.160 0.248 0.504 0.141 -0.037 -0.052 Loss 0.314 0.464 0.000 0.467 0.499 0.000 -0.153 <0.001 CfoVolit 0.059 0.154 0.031 0.094 0.260 0.037 -0.034 <0.001 SalesVol 0.187 0.343 0.115 0.216 0.521 0.105 -0.029 <0.001 %∆CashSales 0.123 0.551 0.065 0.205 1.154 0.054 -0.081 <0.001 ∆ROA 0.002 0.131 0.000 0.006 0.255 -0.001 -0.004 0.065 E/P -0.128 1.244 0.033 -0.300 1.664 0.009 0.172 <0.001 CEOChair 0.400 0.490 0.000 0.273 0.446 0.000 0.127 <0.001 LnAnalyst 6.255 7.233 3.917 3.871 6.036 1.000 2.383 <0.001 ACSize 4.041 1.150 4.000 3.851 1.093 4.000 0.190 <0.001 Independence 0.665 0.163 0.667 0.668 0.187 0.667 -0.003 <0.001 ACExperts 0.370 0.314 0.333 0.265 0.296 0.250 0.105 <0.001 Big4 0.761 0.426 1.000 0.641 0.480 1.000 0.120 <0.001 LnAudtenure 1.671 1.101 1.386 1.400 0.911 0.693 0.272 <0.001 %Loss 0.303 0.358 0.250 0.434 0.387 0.500 -0.131 <0.001 SalesGrowth 0.925 7.913 0.204 2.382 16.057 0.153 -1.458 <0.001 Inventory 0.125 0.124 0.097 0.106 0.133 0.055 0.019 <0.001 lnAuditFees 13.791 1.466 13.875 13.009 1.417 13.016 0.782 <0.001 ∆Inventory 0.003 0.021 0.000 0.002 0.023 0.000 0.001 <0.001 ∆Receivables 0.004 0.028 0.003 0.004 0.034 0.002 0.001 0.015 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 3 (continued) st th All variables are defined in Appendix A. Continuous variables are winsorized at the1 and 99 percentile each year. N= 47,276 overall, N=13,337 for Government contractors, and N=33,871 for non-Government contractors. Internal control weakness data coverage start in 2004. Therefore, there are fewer observations for internal control weakness tests. N=44,410 for C-Score (N=13,708 for Government contractors and N=30,702 for non- Government contractors), 54,420 for REM (N=15,364 for Government contractors and N=39,056 for non-Government contractors), and 18,698 for LnARC measures (N=5,352 for Government contractors and N=13,346 for non-Government contractors). LnARC is available for substanial fewer observations because the financial complexity data coverage starts in 2009. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 4 Pairwise Correlations Panel A: Variables 1 to 10 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) 1.00 (2) 0.03* 1.00 (3) -0.01* -0.02* 1.00 (4) 0.05* -0.02* 0.21* 1.00 (5) -0.10* -0.02* 0.00 -0.01* 1.00 (6) -0.10* -0.02* 0.00 -0.01* 0.99* 1.00 (7) -0.32* -0.05* 0.04* -0.01 0.20* 0.20* 1.00 (8) 0.23* 0.02* -0.01* 0.00 -0.04* -0.04* -0.18* 1.00 (9) -0.06* -0.01 -0.03* -0.01 0.14* 0.14* 0.15* -0.05* 1.00 (10) 0.16* 0.00 0.00 0.02* -0.04* -0.03* -0.05* 0.22* 0.02* 1.00 (11) 0.24* 0.00 0.02* 0.02* -0.14* -0.13* -0.42* 0.11* -0.15* 0.10* (12) 0.17* 0.04* -0.02* 0.00 -0.06* -0.06* -0.31* 0.16* -0.01 0.20* (13) 0.05* 0.01 -0.01 0.00 -0.03* -0.03* -0.20* 0.01* 0.02* 0.06* (14) 0.10* 0.00 0.00 0.01* -0.04* -0.03* -0.04* 0.03* -0.06* -0.01 (15) 0.03* 0.15* -0.01* -0.01 -0.01 -0.01 -0.03* -0.02* -0.02* -0.04* (16) -0.10* 0.06* -0.01 -0.01* 0.05* 0.05* 0.11* 0.01 0.03* -0.17* (17) -0.10* 0.00 0.01 0.01 0.12* 0.12* 0.22* -0.03* 0.08* -0.04* (18) -0.21* -0.04* 0.00 -0.03* 0.19* 0.19* 0.64* -0.04* 0.05* -0.06* (19) -0.17* -0.02* 0.03* 0.01 0.20* 0.20* 0.41* -0.07* 0.16* -0.09* (20) -0.14* -0.01* 0.00 0.00 0.15* 0.14* 0.33* -0.05* 0.10* -0.06* (21) -0.15* -0.01* 0.02* 0.00 0.17* 0.16* 0.28* -0.05* 0.09* -0.11* (22) -0.11* -0.01* 0.00 -0.03* 0.05* 0.05* 0.26* -0.05* -0.03* -0.04* (23) -0.14* -0.02* 0.00 0.00 0.14* 0.13* 0.32* -0.06* 0.21* -0.03* (24) 0.26* 0.02* 0.01* 0.02* -0.15* -0.15* -0.50* 0.13* -0.15* 0.11* (25) 0.06* 0.01 0.00 0.02* -0.05* -0.04* -0.04* 0.02* -0.06* 0.01 (26) -0.07* 0.02* 0.00 0.00 0.07* 0.06* -0.04* -0.07* 0.12* 0.01* (27) -0.24* -0.05* 0.10* 0.02* 0.24* 0.23* 0.86* -0.12* 0.15* -0.02* (28) 0.03* 0.24* -0.01 0.01 0.02* 0.01* 0.06* 0.02* -0.01* -0.03* (29) 0.06* 0.26* 0.00 0.00 0.01 0.01 0.05* 0.03* -0.03* -0.04* Electronic copy available at: https://ssrn.com/abstract=3807526 Table 4 (continued) Panel B: Variables 11 to 20 (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (11) 1.00 (12) 0.20* 1.00 (13) 0.09* 0.46* 1.00 (14) 0.08* -0.03* -0.07* 1.00 (15) 0.13* -0.04* -0.03* 0.08* 1.00 (16) -0.17* -0.11* -0.10* 0.03* 0.11* 1.00 (17) -0.16* -0.08* -0.05* -0.01 -0.01* 0.07* 1.00 (18) -0.30* -0.18* -0.15* -0.01* -0.02* 0.13* 0.22* 1.00 (19) -0.19* -0.14* -0.10* -0.02* -0.02* 0.12* 0.43* 0.41* 1.00 (20) -0.15* -0.11* -0.09* -0.02* -0.01* 0.10* 0.28* 0.35* 0.60* 1.00 (21) -0.11* -0.12* -0.11* -0.01 -0.01* 0.12* 0.40* 0.33* 0.85* 0.66* (22) -0.09* -0.08* -0.05* -0.01 -0.01* 0.02* 0.06* 0.21* 0.10* 0.09* (23) -0.17* -0.07* -0.04* -0.05* -0.02* 0.06* 0.11* 0.25* 0.24* 0.17* (24) 0.77* 0.27* 0.13* 0.08* 0.08* -0.16* -0.17* -0.35* -0.21* -0.16* (25) 0.06* -0.02* -0.04* 0.15* 0.03* -0.02* -0.02* -0.03* -0.04* -0.04* (26) -0.13* -0.03* 0.06* -0.05* -0.02* 0.00 0.01* -0.08* 0.01* 0.01 (27) -0.30* -0.20* -0.14* -0.04* -0.03* 0.08* 0.21* 0.59* 0.45* 0.39* (28) -0.12* -0.09* -0.11* 0.13* 0.04* 0.11* 0.04* 0.05* 0.04* 0.04* (29) -0.06* -0.11* -0.13* 0.15* 0.09* 0.13* 0.04* 0.05* 0.03* 0.03* Panel C: Variables 21 to 29 (21) (22) (23) (24) (25) (26) (27) (28) (29) (21) 1.00 (22) 0.08* 1.00 (23) 0.17* 0.09* 1.00 (24) -0.11* -0.10* -0.20* 1.00 (25) -0.03* 0.00 -0.06* 0.08* 1.00 (26) -0.01* -0.03* 0.04* -0.15* -0.04* 1.00 (27) 0.36* 0.25* 0.30* -0.36* -0.04* -0.05* 1.00 (28) 0.05* 0.01 -0.01* -0.08* 0.04* 0.16* 0.03* 1.00 (29) 0.04* 0.00 -0.02* -0.04* 0.05* -0.03* 0.02* 0.27* 1.00 Table 4 shows Pearson correlation coefficients for variables included in Equations: (1) AbsAccruals (2) AccrualQuality (3) ICMW (4) Restatement (5) CorpGov (6) ContractSize (7) lnAsset (8) M/B (9) FirmAge (10) Leverage (11) Loss (12) CfoVolit (13) SalesVol (14) %∆CashSales (15) ∆ROA (16) E/P (17) CEOChair (18) LnAnalyst (19) ACSize (20) ACExperts (21) Independence (22) Big4 (23) LnAudtenure (24) %Loss (25)SalesGrowth (26)Inventory (27) lnAuditFees (28) ∆Inventory (29) ∆Receivables. All variables are defined in the Appendix. * represent significance at the 0.01 level. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 5 Regression of Accruals Management on Government Contracting AbsAccruals (or AccrualQuality) = β + β CORP_GOV + β lnAsset + β M/B + 0 1 2 3 β Leverage + β FirmAge + β Loss + β %ΔCashSales + β ΔROA + β CfoVolit + 4 5 6 7 8 9 β SalesVol + β E/P+ β lnAnalyst + β CEOChair + β ACSize + 10 11 12 13 14 β Independence + β ACExperts + β Big4 + β LnAudtenure + Fixed Effects + ɛ 15 16 17 18 Panel A: Dependent Variable Equals Absolute Value of Abnormal Accruals Model (1) Model (2) Variables Pred Coefficients t-stat. Coefficients t-stat. *** *** Constant 0.213 12.54 0.214 12.62 *** GovContract H1: - -0.009 4.14 *** ContractSize H1: - -0.011 4.09 *** *** lnAsset - -0.020 19.59 -0.020 19.61 *** *** M/B + 0.006 4.94 0.006 4.94 *** *** Leverage ? 0.045 5.09 0.045 5.10 *** *** FirmAge - -0.004 2.91 -0.004 2.91 *** *** Loss + 0.038 13.92 0.038 13.93 *** *** %∆CashSales + 0.017 8.34 0.017 8.34 ∆ROA + 0.020 0.92 0.020 0.92 *** *** CfoVolit + 0.045 3.16 0.045 3.16 * * SalesVol + -0.008 1.85 -0.008 1.85 *** *** E/P ? -0.008 6.65 -0.008 6.64 * * LnAnalyst - -0.002 1.68 -0.002 1.69 *** *** CEOChair + -0.006 3.18 -0.006 3.18 *** *** ACSize ? 0.004 3.53 0.004 3.53 *** *** Independence - -0.048 5.43 -0.048 5.43 *** *** ACExperts - -0.036 10.31 -0.036 10.32 *** *** Big4 - -0.026 8.63 -0.026 8.63 ** ** LnAudtenure ? 0.002 2.18 0.002 2.18 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 18.46% 18.45% Observations 58,972 58,972 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 5 (continued) Panel B: Dependent Variable Equals Accruals Quality Model (1) Model (2) Variables Pred. Coefficients t-stat. Coefficients t-stat. *** *** Constant 0.019 3.97 0.020 4.05 *** GovContract H1: - -0.002 3.04 *** ContractSize H1: - -0.003 2.91 *** *** lnAsset - -0.002 5.31 -0.002 5.33 M/B + 0.000 0.57 0.000 0.57 Leverage ? 0.001 0.64 0.001 0.65 *** *** FirmAge - -0.001 2.65 -0.001 2.65 *** *** Loss + -0.009 9.79 -0.009 9.79 ** ** %∆CashSales + -0.001 2.14 -0.001 2.14 *** *** ∆ROA + 0.064 12.01 0.064 12.01 *** *** CfoVolit + 0.019 3.42 0.019 3.42 SalesVol + -0.002 1.20 -0.002 1.20 *** *** E/P ? 0.003 6.20 0.003 6.20 *** *** LnAnalyst - -0.002 3.91 -0.002 3.92 ** ** CEOChair + 0.001 2.09 0.001 2.09 ACSize ? 0.001 1.46 0.001 1.47 Independence - -0.003 1.24 -0.003 1.24 ACExperts - 0.001 0.84 0.001 0.83 Big4 - -0.001 0.46 -0.001 0.46 LnAudtenure ? -0.001 1.06 -0.001 1.06 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 3.24% 3.25% Observations 58,972 58,972 Table 5 reports results from estimating Equation (3) using OLS. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 6 Regression of Internal Control Quality on Government Contracting Prob ( ICMW = 1) = F [ δ + δ CORP_GOV + δ lnAsset + δ M/B +δ Leverage + 0 1 2 3 4 δ FirmAge + δ %Loss + δ SalesGrowth + δ Inventories + δ LnAuditFee+ 5 6 7 8 9 δ LnAnalyst + δ CEOChair + δ ACSize + δ Independence + δ ACExperts + 10 11 12 13 14 δ Big4 + δ LnAudtenure + Fixed effects + ɛ ] 15 16 Model (1) Model (2) Variables Pred. Coefficients z-stat. Coefficients z-stat. *** *** Constant 16.842 17.55 16.809 17.51 *** GovContract H1: - -0.260 3.65 *** ContractSize H1: - -0.309 3.48 *** *** lnAsset - -0.502 11.79 -0.503 11.81 M/B - -0.026 1.21 -0.026 1.21 * * Leverage + -0.168 1.88 -0.167 1.88 *** *** FirmAge - -0.234 5.98 -0.235 5.98 ** ** %Loss + 0.204 2.17 0.204 2.17 SalesGrowth + 0.002 1.25 0.002 1.27 Inventory + 0.121 0.37 0.122 0.38 *** *** lnAuditFees + 1.288 20.27 1.288 20.28 LnAnalyst - -0.057 1.54 -0.058 1.56 CEOChair + 0.047 0.79 0.048 0.80 * * ACSize ? -0.050 1.77 -0.049 1.76 *** *** Independence - -0.349 3.26 -0.350 3.27 ACExperts - 0.063 0.32 0.064 0.33 *** *** Big4 - -0.665 7.56 -0.665 7.57 *** *** LnAudtenure ? 0.285 10.03 0.285 10.02 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Pseudo R 11.20% 11.10% Observations 47,267 47,267 Table 6 reports results from estimating Equation (4) using Logit. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 7 Regression of Accounting Restatements on Government Contracting Prob (Restatement =1) = F [δ + δ CORP_GOV + δ lnAsset + δ M/B + δ Leverage + 0 1 2 3 4 δ FirmAge + δ ΔInventories + δ ΔReceivables + δ LnAuditFee + δ %ΔCashSales + 5 6 7 8 9 δ E/P δ ΔROA + δ LnAnalyst + δ CEOChair + δ ACSize + δ Independence + 10 + 11 12 13 14 15 δ ACExperts + δ Big4 + δ LnAudtenure + Fixed effects + ɛ ] 16 17 18 Model (1) Model (2) Variables Pred. Coefficients z-stat. Coefficients z-stat. *** *** Constant 6.013 11.66 5.986 11.61 *** GovContract H1: - -0.199 4.69 *** ContractSize H1: - -0.235 4.48 *** *** lnAsset - -0.142 7.41 -0.143 7.44 M/B - -0.001 0.16 -0.001 0.17 ** ** Leverage + 0.055 2.09 0.055 2.10 *** *** FirmAge - -0.185 7.55 -0.185 7.55 ∆Inventory + 1.079 1.40 1.078 1.40 ∆Receivables + 0.392 0.75 0.394 0.75 *** *** lnAuditFees + 0.275 8.85 0.276 8.85 *** *** %∆CashSales + 0.046 3.27 0.046 3.28 *** *** E/P + -0.024 3.16 -0.024 3.15 ∆ROA - -0.124 1.57 -0.124 1.57 *** *** LnAnalyst - -0.107 4.84 -0.108 4.86 CEOChair ? -0.023 0.62 -0.023 0.61 ACSize ? 0.006 0.38 0.006 0.38 Independence ? -0.104 1.04 -0.104 1.04 ** ** ACExperts - -0.153 2.40 -0.154 2.42 *** *** Big4 - -0.447 9.39 -0.447 9.39 *** *** LnAudtenure ? 0.621 34.63 0.621 34.61 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Pseudo R 7.92% 7.91% Observations 58,972 58,972 Table 7 reports results from estimating Equation (5) using Logit. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 8 Regression of Change in Discretionary Accruals on Change in Government Contract Size ∆AbsAccruals ∆AccrualQuality Variables Coefficients t-stat. Coefficients t-stat. Constant -0.041 1.91 -0.013 1.15 ** ** ∆ContractSize -0.018 2.17 -0.009 2.04 *** ∆lnAsset 0.094 5.13 -0.006 0.67 ∆M/B 0.002 0.60 0.000 0.07 ∆Leverage 0.000 0.21 -0.000 1.28 ** FirmAge -0.092 2.52 0.022 0.90 Loss 0.009 1.14 0.000 0.05 %∆CashSales -0.053 1.11 0.060 1.40 ∆ROA 0.009 0.50 -0.002 0.18 *** ∆CfoVolit 0.004 0.59 -0.010 3.51 ∆SalesVol 0.037 0.48 0.026 1.18 ** ∆E/P 0.006 2.24 0.000 0.34 ∆LnAnalyst 0.000 0.03 0.005 1.81 *** CEOChair -0.002 3.93 -0.000 0.07 ACSize 0.002 0.87 0.000 0.10 Independence 0.002 0.12 0.007 0.69 ACExperts -0.001 0.17 -0.001 0.13 Big4 0.004 0.35 -0.005 0.84 LnAudtenure 0.001 0.13 -0.001 0.36 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 14.60% 8.86% Observations 14,576 14,576 Table 8 report results from estimating a change specification of Equation (3) using only government contractors to provide evidence on the importance of government contract size. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 9 Alternative Financial Reporting Quality Measures Model (1) Model (2) Model (3) C_Score REM LnARC Variable Coefficients t-stat. Coefficients t-stat. Coefficients t-stat. *** *** *** Constant 0.420 38.86 0.347 5.24 3.407 48.68 ** ** ** GovContract 0.002 2.00 -0.028 2.47 -0.019 2.10 *** *** *** lnAsset -0.044 105.95 -0.084 0.162 44.60 16.92 *** M/B 0.047 26.86 -0.004 1.61 -0.005 1.33 *** *** *** Leverage 0.068 17.12 0.156 3.97 0.132 8.30 *** *** FirmAge 0.000 0.51 -0.019 2.71 -0.028 5.46 *** *** *** ∆Inventory 0.010 7.83 0.229 0.056 7.03 19.27 *** *** ∆Receivables -0.003 5.74 0.050 4.95 0.001 0.36 ** ** lnAuditFees -0.016 2.55 0.207 2.28 0.005 0.52 *** *** *** %∆CashSales -0.045 6.87 0.348 5.45 0.063 2.84 *** *** E/P 0.008 3.89 0.026 1.15 0.058 4.37 * *** ∆ROA -0.005 1.68 -0.000 0.00 -0.026 5.31 *** *** LnAnalyst -0.011 17.79 0.006 0.93 -0.059 10.00 *** * *** CEOChair 0.007 6.86 0.020 1.83 -0.037 4.53 *** * ACSize -0.000 0.45 0.018 3.47 0.007 1.89 *** *** Independence -0.006 3.39 -0.002 0.12 -0.137 9.17 *** ACExperts -0.015 5.19 -0.056 1.52 0.005 0.17 *** ** Big4 -0.014 10.79 0.016 1.07 -0.025 2.06 *** ** *** LnAudtenure 0.001 2.60 0.011 1.96 0.016 3.70 Year Fixed Effects Yes Yes Yes Industry Fixed Effects Yes Yes Yes Adjusted R 0.644 0.144 0.599 Observations 44,409 54,134 18,590 Table 9 reports results from estimating Equation (3) using alternative financial reporting quality measures and OLS. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 10 Analysis of Government Contract Gains and Losses Model (1) Model (2) ∆AbsAccruals ∆AccrualQuality Variables Coefficients t-stat. Coefficients t-stat. *** *** Constant -0.025 4.20 -0.012 3.70 *** * GovContractLoss 0.020 2.90 0.007 1.67 t-1 * ** GovContractGain -0.014 1.73 -0.008 2.25 t-1 *** ∆lnAsset 0.038 7.76 -0.000 0.15 * * ∆M/B 0.004 1.92 0.002 1.82 *** ∆Leverage 0.000 3.53 0.000 0.99 *** FirmAge -0.109 3.18 0.012 0.96 *** *** Loss 0.037 9.24 0.007 3.77 *** *** %∆CashSales -0.141 3.37 0.061 3.78 ∆ROA -0.016 1.54 -0.003 0.47 ** ∆CfoVolit 0.001 0.19 -0.005 2.57 *** *** ∆SalesVol -0.609 9.78 0.146 5.94 ∆E/P 0.002 0.86 0.002 1.14 ∆LnAnalyst 0.001 0.31 -0.000 0.47 *** ** CEOChair -0.003 4.90 0.001 2.55 ACSize 0.000 0.17 -0.001 0.78 ACExperts -0.003 0.30 -0.005 1.10 Independence 0.002 0.10 0.001 0.15 ** Big4 0.003 0.91 0.004 2.55 *** LnAudtenure 0.003 3.46 -0.001 0.07 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 18.50% 5.40% Observations 29,668 29,668 Table 10 report results from estimating a change specification of Equation (3) using the PSM sample and controlling for gains and losses in government contracts. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png ARN Conferences & Meetings SSRN http://www.deepdyve.com/lp/ssrn/federal-government-contracts-and-financial-reporting-quality-gj0HnOr8X4

Loading next page...

References (19)

Patricia Dechow, Weili Ge, Chad Larson, Richard Sloan (2011)
Predicting Material Accounting Misstatements*: Predicting Material Accounting Misstatements
Contemporary Accounting Research, 28
D. Cohen, Bin Li (2016)
Why Do Firms Hold Less Cash? A Customer Base Explanation
Financial Accounting eJournal
K. Raman, Husayn Shahrur (2008)
Relationship-Specific Investments and Earnings Management: Evidence on Corporate Suppliers and Customers
Financial Accounting
Annie McGowan, Valaria Vendrzyk (2002)
The Relation between Cost Shifting and Segment Profitability in the Defense‐Contracting Industry
The Accounting Review, 77
Kin Lo, Felipe Ramos, Rafael Rogo (2016)
Earnings Management and Annual Report Readability
CGN: Disclosure & Accounting Decisions (Topic)
M. McNichols (2000)
Research design issues in earnings management studies
Journal of Accounting and Public Policy, 19
Daniel Cohen, Bin Li, Ningzhong Li, Yun Lou (2016)
Major government customers and loan contract terms
Review of Accounting Studies, 27
Feng Li (2008)
Annual Report Readability, Current Earnings, and Earnings Persistence
Journal of Accounting and Economics, 45
Hui Chen, Katherine Gunny (2014)
Profitability and Cost Shifting in Government Procurement Contracts
Corporate Social Responsibility (CSR) eJournal
Agnes Cheng, Wenli Huang, Shaojun Zhang (2020)
Major government customer and management earnings forecasts
Frontiers of Business Research in China, 14
M. McNichols (2002)
Discussion of the Quality of Accruals and Earnings: The Role of Accrual Estimation Errors
Accounting review: A quarterly journal of the American Accounting Association, 77
Lillian Mills, Sarah Nutter, Casey Schwab (2011)
The Effect of Political Sensitivity and Bargaining Power on Taxes: Evidence from Federal Contractors
McCombs: Accounting (Topic)
D. Cohen, Bin Li (2019)
Customer-Base Concentration, Investment, and Profitability: The U.S. Government as a Major Customer
The Accounting Review
D. Cohen, Aiyesha Dey, Thomas Lys (2004)
Real and Accrual-Based Earnings Management in the Pre- and Post-Sarbanes Oxley Periods
SPGMI: Compustat Fundamentals (Topic)
Lauren Cohen, Andrea Frazzini (2007)
Economic Links and Predictable Returns
SPGMI: Compustat Fundamentals (Topic)
D. Larcker, Scott Richardson, A. Tuna (2007)
Corporate Governance, Accounting Outcomes, and Organizational Performance
Capital Markets: Asset Pricing & Valuation eJournal
Patricia Dechow, Ilia Dichev (2002)
The Quality of Accruals and Earnings: The Role of Accrual Estimation Errors
The Accounting Review, 77
Grace Pownall (1986)
An Empirical Analysis of the Regulation of the Defense Contracting Industry: The Cost Accounting Standards Board
Journal of Accounting Research, 24
(1992)
A perspective on accounting for defense contracts

Copyright: Copyright © 2021 Elsevier Inc.
eISSN: ARCH-4059
DOI: 10.2139/ssrn.3807526
Publisher site: See Article on Publisher Site

Abstract

Government spending is essential for the US economy, and the amount of capital that flows from the government to US firms increased substantially in recent years. Despite the economic importance of the corporate-government contracting relationship, we know little about the firm- level financial outcomes associated with government contracts. Federal government regulations require federal contractors to maintain strong internal control over financial reporting and government contractors have strong incentives to maintain government contracts. As a result, we expect and find that corporate government contracting relationships are associated with higher firm-level financial reporting quality compared to non-government contractors. Further, we find that the improvement of financial reporting quality begins when a firm becomes a federal contractor and greater amounts of government contracts revenue are associated with higher levels of financial reporting quality. We also find the quality of financial reporting weakens after a firm loses government contracts. Collectively, our empirical results suggest that having the government as a customer has a positive impact on the quality of financial reports. Keywords: public procurement, government contractors, financial reporting quality Data Availability: Data are available from the public sources cited in the text. Electronic copy available at: https://ssrn.com/abstract=3807526 Federal Government Contracts and Financial Reporting Quality 1. Introducation US government spending was $4.45 trillion in 2019, representing 21% of the US Gross Domestic Product. This massive amount of spending drives large portions of the US economy and creates millions of jobs for US citizens (Davidson 2018). While government spending is an essential and substantial component of the US economy, we know relatively little about the effect of the government procurement process on the quality of government suppliers’ firm-level financial reporting. It is important to understand federal government contractor’s financial reporting quality since taxpayers are interested in knowing whether public resources are being managed efficiently and effectively. As a result, we investigate whether supplier-customer contracting relationships between corporations and the US Government affect firm-level financial reporting quality. For firms that have contracts with the US government, the contracts account for a large share of the contractors’ revenue. Ellis, Fee, and Thomas (2012) report that 13.4% of public firms' major customers are the US Government and its agencies accounting for one-third of their respective total revenue on average. As government contracts provide contractors a steady stream of revenue that lowers suppliers’ demand uncertainties and increases profitability (Cohen and Li 2019); however, the contractor firms also become increasingly more government- The US federal government awards about $500 billion in contracts annually and is the largest buyer of goods and services in the nation (Samuels 2021). US federal government expenditure average 20% of US GDP over the last 40 years (Cohen and Li 2019). Electronic copy available at: https://ssrn.com/abstract=3807526 dependent (Ellis et al. 2012). Corporate managers therefore have strong incentives to maintain these corporate-government contracting relationships in order to enhance firms’ profitability, improve their stock valuation, and lower their cost of capital (e.g., Cohen, and Li 2019, Banerjee, Dasgupta, and Kim 2008, and Dhaliwal, Judd, Serfling, and Shaikh 2016). Given this incentive, government suppliers provide more voluntary disclosures relative to non-government contractors (Samuels 2021). Existing literature provides evidence as to how major customers affect firm-level financial reporting quality. For example, Raman and Shahrur (2008) provide evidence that firms use earnings management to influence their major commercial suppliers’ or customers’ perceptions of their business prospects. We expect that government contracts impact firms’ reporting outcomes significantly different from commercial contracts (Cohen and Li 2019). We extend this line of literature by investigating whether being a government contractor are associated with the quality of their financial reports provided to the public. We believe that business transactions with the US Government are associated with higher quality financial reporting by government suppliers for a number of reasons. First, the government supplier faces less competition than the typical commercial supplier because of noncompetitive procurement contracts (Mills, Nutter, and Schwab 2013). The unique nature of government sales motivates government suppliers to increase their relationship-specific Government supplier and governement contractor are used interchangably throughout this dissertation to mean a firm that sells goods or services to the government.Further, US Government refers to all departments and agencies of the US Federal Governement. Electronic copy available at: https://ssrn.com/abstract=3807526 investments in order to maximize transaction efficiencies to ensure a competitive position (Cohen and Frazzini 2008) and thereby enhance profitability (Cohen and Li 2019). As a result, corporate managers are less likely to choose an accounting method that involves shifting income from one period to another, resulting in higher financial reporting quality. Second, government agencies have a low default risk and often offer longer-term procurement contracts relative to commercial contracts (Dhaliwal et al. 2016). Low operational and demand uncertainties enhance the forecastability of corporate outcomes and thereby lower corporate managers’ opportunism to manage earnings to meet personal objectives, reducing agency conflicts between managers and shareholders. Third, the government procurement process is administered according to the Federal Acquisition Regulation (FAR) and the government can terminate a contract at any given time (FAR 49.101). Thus, the government agency has greater bargaining and enforcement power to obtain necessary information from its contractors than private-sector customers (Chaney, Faccio, and Parsley 2011). Government contractors will therefore have a strong incentive to ensure that financial reports have the quality needed to mitigate the potential risk of losing government contracts. We regress measures of financial reporting quality on constructs for corporate- government relationships and control variables based on prior research to test our hypothesis. We find a negative association between both the size and existence of government contracts and the level of discretionary accruals as well as the probability of material weakness in internal control, and restatements, supporting our hypothesis. Our study adds to the literature on implications of corporate-government contracting. We extend Samuels (2021) that focuses on whether federal contractors are more likely to issue Electronic copy available at: https://ssrn.com/abstract=3807526 management forecasts and the frequency of the issuance. Our study, on the other hand, provides extensive evidence on an important aspect of financial accounting, that of reporting quality. Also, our study adds to the stream of research about the effects of bilateral relationships on firm internal decisions and performance, especially studies looking at the consequences of major customer dependency (e.g., Kalwani and Narayandas 1995; Banerjee et al. 2008; Gosman and Kohlbeck 2009). Additionally, our study responds to Healy and Wahlen’s (1999) call for future research to further explain how business factors drive accruals by documenting how firms contracting with the federal government and its agencies exhibit a lower level of accruals management. Our findings should be of interest to investors, as they seek to minimize the risk of investing in government contractors' stocks and make strategic decisions regarding their buy and hold decisions. The findings should also be of interest to regulators (e.g., the Securities and Exchange Commission [SEC]) and other stakeholders (e.g., customers and suppliers) who are interested in understanding the factors that can improve firm’s financial reporting quality, as well as stakeholder groups who care about firms’ future growth potential. Finally, our study adds to financial reporting literature by investigating the financial reporting quality of government contractors. Early studies focus on the relationship between government contracting regulation and contractor firm’s accounting outcome (e.g., Pownall 1986; Horwitz and Normolle 1988). Subsequent studies investigate whether government contractors enjoy more subsidies from federal agencies (Callahan, Vendrzyk, and Butler 2012; Cohen and Li 2016); are affected by political sensitivities that influence firm outcomes (e.g., Karpoff, Lee, and Vendrzyk 1999; Mills, Nutter, and Schwab 2013); or shift costs from commercial to government contracts to extract benefits (e.g., Lichtenberg 1992; McGowan and Electronic copy available at: https://ssrn.com/abstract=3807526 Vendrzyk 2002; Chen and Gunny 2014). Our study therefore adds to the literature by investigating the financial reporting quality of government contractors. We organize the study as follows. The next section provides a literature review and hypothesis development. The third section contains research design and variable descriptions. The fourth section provides a discussion of our sample and descriptive statistics. We report the results of tests in the fifth section, which includes a discussion of our sensitivity analyses. The final section provides a conclusion. 2. Literature Review and Hypothesis Development Corporate-Government Contracting Relationships and Firm Outcomes Early studies investigating the profitability of government contractors find that these contractors earned higher profits from government contracts. However, these studies provide mixed evidence as to why government contractors are able to generate abnormally high profits. For instance, Reichelstein's (1992) findings suggest that government contractors earn an abnormally high profit from their government business by shifting and charging overhead and pension costs from commercial operations to their government contracts. Lichtenberg (1992) confirms these findings and further suggests that government contractors earn almost three times more profit than commercial contractors. His results indicate that government contractors are also significantly less capital-intensive than non-government contractors. However, McGowan and Vendrzyk (2002) find no evidence that government contractors’ abnormal profits are attributable to cost-shifting. Their study finds that abnormal profits result from price renegotiation with the government agency or cost reimbursement. McGowan and Vendrzyk (2002) argue that the unusually high profitability of government contractors is more likely attributable to non-accounting explanations, such as the lack of industry competition. Electronic copy available at: https://ssrn.com/abstract=3807526 Recent literature investigates factors that can influence a government agency’s decisions to grant contracts to firms. Flammer (2018) investigates the association between corporate social responsibility (CSR) and the likelihood of gaining government procurement contracts. The author argues that a prospective contractor uses CSR reporting to signal it’s quality and differentiate itself from other competitors in order to influence a government agency’s purchasing decision. His results indicate that CSR firms receive more government contracts when the contractor is in a competitive industry. However, firms that make larger political contributions are more likely to receive government contracts with favorable terms (Ferris, Houston, and Javakhadze 2019). Other research investigates the effect of the corporate-government contracting relationships on corporate outcomes. This research generally finds that firms benefit from having a government agency as their major customer because investors and creditors view these firms as having a lower risk than firms without government contractors. Government contractors hold significantly less cash and have less volatile future earnings than non-governmental contractors (Cohen and Li 2016). Government contractors also enjoy a lower number of debt covenants and are less likely to have performance pricing provisions in their loan contracts (Cohen, Li, Li, and Lou 2016). More importantly, government contractors tend to have lower costs of equity capital (Dhaliwal et al. 2016). Cohen and Li (2019) find that firm profitability increases with a concentration of major government customers but decreases with a concentration of major corporate customers. The authors conclude that major government customers have unique and significant influences on firm outcome which differs from major commercial customers. Other research suggest that the government plays a useful role in the contract process by monitoring its contractors. Samuels (2021) investigates whether government monitoring of the Electronic copy available at: https://ssrn.com/abstract=3807526 contractor’s internal information process strengthens the firm’s external reporting quality. Examining the frequency of voluntary disclosures, speed of earnings announcements, and quality of public information about the contractor, Samuels suggests that having both a government contract and a high dollar government contract are associated with the contractors having higher quality of external reporting. Moreover, she finds that external reporting quality improves when a firm is granted a government contract compared to firms that are not awarded government contracts. Cheng et al. (2019) find that government contractors issue more precise and accurate management earnings forecasts than non-government contractors. Their results suggest that government agencies have greater bargaining and enforcement power, which enables them to obtain more information about the supplier firm. Hypothesis Development Government contractors are required to submit annual audited financial statements and compliance reports, including a single audit report (CFR 200.331). FAR requrie federal contractors maintain internal control systems that are suitable for the size of the organization, able to timely discover and disclose control deficiencies, and ensure correction actions are carried out (FAR 3.1). Furthermore, the contractor’s independent auditor has to identify whether there is a material weakness in internal control, any instance of material non-compliance with the Federal Code, any questionable costs related to the federal contract and/or, large and unusual amounts of federal contract-related expenses in comparison to total operating costs. As a result of the additional requirements stipulated by the regulations, the corporate-government contracting relationship is likely to be associated with higher financial reporting quality. Major customers, including the federal government, have the ability and incentive to monitor their key suppliers in order to mitigate potential agency costs (Itzkowitz 2013, and Electronic copy available at: https://ssrn.com/abstract=3807526 Itzkowitz 2015). FAR regulations extensively formalize government monitoring policies and procedures for government agencies. These policies and procedures enhance government agency’s abilities to monitor contractors’ internal control and information processing. As a result, effective government monitoring deters earnings management as the government can terminate a contract at any given time if it identifies corporate wrongdoing (FAR 49.101). Government agencies also constantly update their policies and procedures to ensure the monitoring is adequate. Further, according to the positive accounting theory, firms will choose accounting policies that minimize contract costs (Watts and Zimmerman’s 1978). The political cost hypothesis therefore predicts that firms subject to government scrutiny such as government contractors take actions to minimize the probability of negative government actions. Corporate- government contracting is expected to be associated with higher firm-level financial reporting quality. Competition for government contracts is generally less than typical commercial transactions (Mills, Nutter, and Schwab 2013). Government contractors are therefore motivated to make relationship-specific investments to maximize transaction efficiencies. As a result, government contractors enjoy a stronger competitive position (Cohen and Frazzini 2008) and experience enhanced profitability (Cohen and Li 2019). Therefore, there are lower incentives to choose accounting methods that shift income or otherwise lower financial reporting quality. According to CFR 200.331, government contractors are requried to disclose their financial information in order to maintain their government contracts. Electronic copy available at: https://ssrn.com/abstract=3807526 Consistent with this argument, recent studies show that government agencies positively impact firm external reporting as management earnings forecasts are more precise and accurate (Cheng et al. 2019), information asymmetry is reduced, and the firm overall transparency is enhanced (Samuels 2021). Government contractors also benefit from the predictability of operating cash flows from government contracts, the lower default risk, and long-term nature of the contracts. Dechow and Dichev (2002) show that a higher certainty of future cash flows is associated with a higher accrual quality. Accrual quality is also positively associated with earnings persistence. Therefore, government contractors are less likely to smooth earnings in order to signal private information about future performance to stakeholders (Watts and Zimmerman 1986, Sankar and Subramanyam 2001; Tucker and Zarowin 2006). Based on the above arguments, our hypothesis predicts that corporate-government contracting relationships enhance firm-level financial reporting quality. Our hypothesis is stated as follows: HYPOTHESIS: Corporate-government relationships are associated with higher firm-level financial reporting quality. However, a negative association is possible given that firms dealing with government agencies are incentivized to disclose optimistic financial reports. Agency theory suggests that self-interested contracting parties have the incentive to behave opportunistically in situations of uncertainty and information asymmetry (e.g., Jensen and Meckling 1976 and Jensen 1986). Consistent with the earnings management literature (e.g., Raman and Shahrur 2008), a government contractor might use accruals to inflate reported earnings in order to signal to government agencies that it is able to continue to meet its contractual obligations. Electronic copy available at: https://ssrn.com/abstract=3807526 3. Research Design Corporate-Government Contracting Variables We use the Federal Procurement Data System (FPDS) for data on government contractors. The FPDS contains information on all government contracts and contract modifications beginning in 2000. The US government discloses detailed information on government contracts that are above $25,000, including the total contract value, information about the contractor, and the federal agency. Using this data, we identify firms that have a corporate-government contracting relationship in a given year (GovContract). We also use ContractSize, a continuous measure of contract award size relative to a firm’s total sales (e.g., Mills, Nutter and Schwab 2013; Samuels 2021). The impact of the corporate government contracting relationship on the contractor’s financial reporting quality might vary with the importance of the government contracts to the contractor. Because ContractSize is heavily right-skewed, we follow Samuels (2021) and transform the variable into decile ranks scaled from 0 to 1. This transformation has the advantage of being robust to both outliers and nonlinearities and eases the interpretation of the results. The Federal Procurement Data System data is available at USAspending.gov. The estimated regression coefficient measures the change in the respective measure of financial reporting quality when moving from the bottom decile to the top decile of contract size, ceteris paribus). Electronic copy available at: https://ssrn.com/abstract=3807526 Financial Reporting Quality Variables The first construct for financial reporting quality is discretionary accruals. We follow Kothari, Leone, and Wasley’s (2005) approach for performance-matched discretionary accruals measure. This performance-matched discretionary accrual model adjusts for industry and year performance. We first construct a matched sample based on the firm’s closest return on assets (ROA) in the same year and then estimate the following model annually in each three-digit SIC industry with more than 15 observations. TA , /A = β (1/A𝑖 𝑡 ) + β (∆REV𝑖 𝑡 /A 𝑖 𝑡 − ∆REC𝑖 𝑡 /A 𝑖 𝑡 ) + β (PPE𝑖 𝑡 /At−1) + i t i,t-1 1 , −1 2 , , −1 , , −1 3 , β (ROA ) + β (MB ) + 𝜀 (1) 4 𝑖 ,𝑡 5 𝑖 ,t 𝑖 ,t where variables are defined as follows: TA is computed as earnings before extraordinary items minus cash flow from operating activities from the statement of cash flows, A represents total assets, ∆REV is the change in revenues, ∆REC is the change in net receivables, PPE represents the gross amount of property, plant, and equipment, ROA is net income before extraordinary items divided by lagged total assets, and MB is the ratio of the book value of net assets divided by the market capitalization. The residual from Equation (1) proxies for discretionary accruals. Prior studies argue that the modified Jones (1991) model does not capture a firms’ earnings management when financial performance is extreme, resulting in potentially incorrect inferences and model misspecification. Building on the modified Jones model and Dechow et al. (1995), Kothari et al. (2005) use a control sample to create performance-matched discretionary accruals in order to identify abnormal earnings management. Electronic copy available at: https://ssrn.com/abstract=3807526 We then take the absolute value of the discretionary accruals (AbsAccruals) to capture a firm’s financial reporting quality. Greater AbsAccruals indicate poorer financial reporting quality. Our next construct for financial reporting quality is the accruals quality measure developed by Dechow and Dichew (2002) as modified by McNichols (2002). Dechow and Dichev (2002) document that working capital accruals are affected by cash flows by mapping the working capital accruals into operating cashflows. McNichols (2002) modifies the measure by including property, plant and equipment, and changes in sales. Following Ball and Shivakumar (2006) and Dou, Khan, and Zou (2016), we estimate the following model annually in each three- digit SIC industry with more than 15 observations: TA = β + β CFO + β CFO + β CFO + β ΔSALES + β PPE + 𝜀 (2) i,t 0 1 𝑖 ,𝑡 -1 2 𝑖 ,𝑡 3 𝑖 ,𝑡 +1 4 𝑖 ,𝑡 5 𝑖 ,𝑡 𝑖 ,𝑡 where variables are defined as follows: TA is as earnings before extraordinary items minus cash flow from operating activities from the statement of cash flows, CFO is cash flow from operating activities, ΔSALES is the year to year change in total sales, and PPE is the gross amount of property, plant, and equipment. All variables are divided by average total assets. The residual from Equation (2) is the accrual quality measure (AccrualQuality). Greater values of AccrualQuality indicate poorer financial reporting quality. A firm’s financial reporting quality is a reflection of its willingness to put a strong internal control structure in place to ensure that financial reports are of adequate quality. Consequently, our third construct for financial reporting quality is the presence of an internal control material weakness. Internal control material weakness data is collected from the Audit Analytics database. We define ICMW as an indicator variable equal to one if there is a general (systemic) material weakness in internal controls reported under either SOX 302 (management) Electronic copy available at: https://ssrn.com/abstract=3807526 or SOX 404 (auditor) in a given year, and zero otherwise. We also consider the number of material weaknesses reported in our sensitivity tests. Restatements capture the failure of a firm’s financial reporting process. Restatement data is obtained from the Audit Analytics database. Restatement is an indicator variable equal to one if the government contractor’s financial statement for that year is subsequently restated, and zero otherwise. Later sensitivity tests consider intentional misstatements as it is important to distinguish error from irregularities (Hennes, Leone, and Miller 2008). This second restatement measure captures restatements that were followed by fraud-related class action lawsuits. Corporate-Government Relationship and Accruals Earnings Management Our first model tests the relationship between corporate-government contracting and financial reporting quality using discretionary accruals and accruals quality. We follow Ham, Lang, Seybert, and Wang (2017) for our empirical specification modified to include governance variables to explain accruals earnings management (subscripts for firm and year are omitted here and in later equations for brevitiy unless necessary to understand the equation). AbsAccruals (or AccrualQuality) = β + β CORP_GOV + β lnAsset + β M/B + 0 1 2 3 β Leverage + β FirmAge + β Loss + β %ΔCashSales + β ΔROA + 4 5 6 7 8 β CfoVolit + β SalesVol + β E/P+ β lnAnalyst + β CEOChair + 9 10 11 12 13 β ACSize + β Independence + β ACExperts + β Big4 + 14 15 16 17 We exclude the CEO characteristics control variables from Ham et al. (2017). Their paper foucses on CFO and CEO characteristics whereas our paper focuses on firm characteristics. Electronic copy available at: https://ssrn.com/abstract=3807526 β LnAudtenure + Fixed Effects + ɛ (3) where variables are defined in Appendix A. Year and two-digit SIC industry fixed effects are included in our estimation to mitigate the impact of time-invariant year and industry characteristics. Standard errors are clustered by firm. As discussed above, we consider two accruals earnings management variables for financial reporting quality - discretionary accruals (AbsAccruals), and accruals quality (AccrualQuality). The key variable of interest is CORP_GOV that is defined as either the presence or amounts of the corporate-government relationship (GovContract and ContractSize). Consistent with our hypothesis, we expect a negative coefficient on both GovContract and ContractSize. Negative coefficients suggests that a government contractor is less likely to engage in accruals earnings management and hence, have higher financial reporting quality. We include several firm characteristics that might influence the level of accrual earnings management. Larger firms have better financial reporting processes and procedures in place to ensure the quality of accruals (e.g., Kinney and McDaniel 1989; Ge and McVay 2005). We expect the coefficients of lnAsset to be negative, indicating that larger firms are associated with higher accruals quality. Market to book ratio and leverage capture growth opportunities and default risk. High growth firms and firms with high default risk are more likely to choose accounting procedures that shift reported earnings from a future period to the current period (e.g., Watts and Zimmerman 1986), resulting in higher discretionary accruals. However, Jensen (1986) suggests the use of debt increase outside monitoring and therefore reduces managers’ opportunitstic behaviors. We expect the coefficient on M/B to be positive but are unable to predict the coefficient on Leverage. Firm age is an indication of firm maturity where younger Electronic copy available at: https://ssrn.com/abstract=3807526 firms are less likely to have established procedures. Thus, we expect the coefficient of FirmAge to be negative, indicating that older firms have higher accruals quality. Firm financial performance influences the level of accruals management. We control for losses, percentage change in cash sales, and change in return-on-assets We expect firms experiencing financial difficulties may pay less attention to the financial reporting process. Greater accrual management generally accompanies increases in ROA (McNichols 2000 and McNichols 2002). Thus, we expect the coefficients on Loss, %ΔCashSales, and ΔROA to be positive. Dechow and Dichev (2002) suggest that cash flows and sales volatilities are significantly associated with accruals. We expect the coefficients of CfoVolit, and SalesVol to be positive, reflecting management’s incentive to smooth earnings for compensation incentives (e.g., Jayaraman 2008, Garrett, Hoitash and Prawitt. 2014). We also control for the earnings to price ratio in order to reflect investors’ anticipated demand for earnings growth in future periods. However, a higher earnings to price ratio might be associated with higher or lower earnings quality (Ham et al. 2017). Firm governance is also a key determinant of a firm's financial reporting quality. We first control for the number of analysts following the firm in order to capture external monitoring. Higher external monitoring should restrict management’s ability to manage earnings, resulting in a higher level of accruals quality. We expect the coefficient of LnAnalyst to be negative. We then control for CEO duality, where the CEO is the chairman of the board. Agency theory argues that powerful CEOs have managerial discretion to manipulate earnings in order to meet personal goals. We expect the coefficient of CEOChair to be positive. Audit committee financial expertise (ACExpert) and independent boards of directors (Independence) enhance board monitoring to ensure the quality of financial reports (e.g., DeFond, Hann, and Hu 2005; Electronic copy available at: https://ssrn.com/abstract=3807526 Dhaliwal, Naiker, and Navissi 2010; Carcello, Hermanson, Neal, and Riley 2002). Thus, we expect the coefficients on ACExpert and Independence to be negative. Audit committee size may also impact earnings management (Klein 2002); however, the empirical evidence is mixed (e.g., Larcker, Richardson, and Tuna 2007). As a result, we do not predict the coefficient of ACSize. We control for Big 4 auditors since prior studies find Big 4 auditors have a significantly positive impact on financial reporting quality (e.g., Carcello et al. 2011). Finally, we include auditor tenure (LnAudtenure); however, we are unable to predict the sign of the coefficient since the impact of auditor tenure on audit quality is uncertain (Singer and Zhang 2018). Corporate-Government Relationship and Internal Control Material Weakness Consistent with FAR requirements, we expect government contractors to have stronger internal control structures that improve the quality of financial reports. For our internal control material weakness model, we start with Ham et al. (2017) and add controls for governance. Our specification is as follows: Prob ( ICMW = 1) = F [ δ + δ CORP_GOV + δ lnAsset + δ M/B +δ Leverage + 0 1 2 3 4 δ FirmAge + δ %Loss + δ SalesGrowth + δ Inventories + δ LnAuditFee + 5 6 7 8 9 δ LnAnalyst + δ CEOChair + δ ACSize + δ Independence + 10 11 12 13 δ ACExperts + δ Big4 + δ LnAudtenure + Fixed effects + ɛ ] (4) 14 15 16 where variables are defined in Appendix A and F is a logistic function. We again include year, and two-digit SIC industry fixed effects in our estimation to mitigate the impact of time-invariant year and industry characteristics. Consistent with our hypothesis, negative coefficients on the two CORP_GOV constructs suggest that government contracts are less likely to receive an ICMW, corresponding to a higher level of financial reporting quality. Electronic copy available at: https://ssrn.com/abstract=3807526 Similar to Equation (3), we include firm size, market-to-book ratio, leverage, and firm age. Prior research indicates that larger, growth, older, and firms with lower leverage are associated with higher financial reporting quality (e.g., Ge and McVay 2005, Watts and Zimmerman 1986), Positive (negative) coefficients are therefore expected for Leverage (LnAssets, M/B and FirmAge). We follow prior studies (e.g., Ham et al. 2017; Krishnan 2005 and Ashbaugh‐Skaife, Collins, and Kinney 2007) and control for firm operating risk and business complexity using the number of years that the firm reports negative earnings over the prior four years, sales growth, and inventory. A risky firm with higher business complexity is associated with a higher probability of internal control weaknesses. Thus, we expect %Loss, SalesGrowth, and Inventory to be positively associated with ICMW. We also use audit fees to control for auditor effort; however, we are unable to predict the sign of audit fee since there is a large array of risk factors that are priced into audit fees and prior research documents mixed results between audit fees and audit quality (DeFond and Zhang 2014). Finally, we include a number of variable to control for firm governance. Stronger governance is associated with a lower probability of internal control material weaknesses. Accordingly, we expect inverse relationships for the number of analysts following the firm, audit committee expertise, board of director independence, and Big 4 auditor and a positive association for CEO – chairman duality. Consistent with prior research, we leave the prediction unsigned for audit committee size and auditor tenure. Corporate Government Relationship and Restatements Our final construct for the quality of financial statements is Restatements. We again follow Ham et al. (2017) for our restatement model. To the extent that firms with government Electronic copy available at: https://ssrn.com/abstract=3807526 contracts have a lower tendency to misreport, we expect to find a decreased likelihood that financial statements are misreported. Firms rarely restate their financial statements, and restatements do not capture all types of misreporting. However, the restatement tests represent a more objective outcome of misreporting. We use the following Logit model for our test: Prob (Restatement =1) = F [δ + δ CORP_GOV + δ lnAsset + δ M/B + δ Leverage + 0 1 2 3 4 δ FirmAge + δ ΔInventories + δ ΔReceivables + δ LnAuditFee + 5 6 7 8 δ %ΔCashSales+ δ E/P δ ΔROA + δ LnAnalyst + δ CEOChair+ 9 10 + 11 12 13 δ ACSize + δ Independence + δ ACExperts + δ Big4 + 14 15 16 17 δ LnAudtenure + Fixed effects + ɛ ] (5) where variables are defined in Appendix A. We include year and two-digit SIC industry fixed effects to mitigate the impact of time-invariant year and industry characteristics. Consistent with our hypothesis, a negative coefficient on our two constructs for CORP_GOV suggests that a government contractor is less likely to restate the financial statements and hence, have higher financial reporting quality. We again include firm size, market-to-book ratio, leverage, and firm age in this financial reporting quality specification. We expect that larger, growth, older, and less levered firms are associated with higher financial reporting quality and therefore a lower probability of restatements. Prior research suggests that an increase in working capital and a higher percentage changes in cash sales is associated with earnings management and a higher probability of future restatements (e.g., Beneish 1999; Dechow, Ge, Larson and Sloan 2011). We therefore include ΔInventories, ΔReceivables, and %ΔCashSales. Audit fees are associated with a greater likelihood of restatements consistent with increased audit risk (e.g, Armstrong, Blouin, Electronic copy available at: https://ssrn.com/abstract=3807526 Jagolinzer and Larcker 2015). We also expect higher earnings to price ratio and lower changes in ROA are associated with restatements (e.g., Dechow et al. 2011). Finally, the governance variables included in the discretionary accruals models are included. As the number of analysts following the firm, the proportion of financial experts on the audit committee increases, or the firm has a Big 4 auditor, we expect the probability of restatement declines. However, we make no prediction for the remaining governance variables. 4. Sample Selection and Descriptive Statistics Sample Selection We start with the cross-section of the CRSP and Compustat datasets to obtain our required financial and accounting data over the period from 2001 to 2017. Following prior studies, we exclude firms in highly regulated industries (banking and utilities), which results in 115,572 firm-year observations. We then combine these observations with our government contractor data we assembled using the FPDS. As the contract data does not contain common identifiers such as GVKEY or PERMNO, we perform a fuzzy-match of the parent company names to the names of Compustat companies, and hand-check identified pairs to ensure proper matching. We aggregate all government contract information by matched firm and year. We eliminate 41,392 and 8,712 firm-year observations with missing data to compute the discretionary accruals and control variables, respectively. Next, we eliminate an additional 6,496 firm-year observations with data missing from Audit Analytics. This provides a sample size of 58,972 firm-year observations to estimate the discretionary accruals and restatement models. Since the internal control material weakness data is only available beginning in 2004, we use the second sample of 47,276 firm-year observations to estimate the internal control material weaknesses models. Table 1 presents the sample construction. Electronic copy available at: https://ssrn.com/abstract=3807526 <Insert Table 1> Table 2, Panel A provides the sample distribution by year. The annual distribution is fairly constant across years between the government contractor and total samples. The total government contracts per year in our sample range from $75 billion in 2001 to $200 billion in 2008. Consistent with prior studies (e.g., Mills, Nutter, and Schwab 2013, and Samuels 2021), the annual government contract value averages about 2 to 4 percentage of total firm sales. <Insert Table 2> Sample distributions by the Fama-French 12 industry classifications are presented in Panel B of Table 2. Our sample covers a wide range of industries with computer and software firms representing the largest industry. The computer and software firms also are the most represented among the government contractor firms. Government contractors represent between 21% and 35% of the number of firms for most industries. In terms of government contract value, manufacturing and computers and software are the largest industries ($1.1 trillion, and $700 billion, respectively). The relative importance of government contracts (as a percentage of total sales) is also the greatest in these two industries at 9.0% and 4.6%, respectively. Descriptive Statistics Panel A of Table 3 presents descriptive statistics for the model variables. The mean value of CorpGov indicates that about 27% of the sample firms are government contractors. On average, the sample firms are characterized as large and profitable indicated by mean total assets of approximately $249 million and mean change of 0.5% in return on assets. The sample firms report total debt averaging 23.8% of total assets and lower inherent risk as inventory represents only 11% of total assets. The average firm is just over ten years old. The mean audit committee size and board independence for the sample firms are 3.9 directors and 66.7%, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Furthermore, 29% of audit committee members have financial expertise, and 67.7% of our sample firms are audited by one of the Big 4 auditors. Overall, our sample statistics are similar to those reported by prior research (e.g., Mills, Nutter, and Schwab 2013, and Samuels 2021). In Panel B of Table 3, we compare firm-year observations with and without government contracts. The univariate tests show that firms with government contracts are characterized by significantly lower discretionary accruals and a lower probability of restatements than non- government suppliers. This significant differences in discretionary accruals and restatement numbers provide initial support for our hypothesis. Government contractors are also significantly larger, more profitable, more mature, stronger corporate governance, and subject to more external monitoring. <Insert Table 3> The Pearson correlation coefficients for the model variables are presented in Table 4. The correlations between discretionary accruals and government contracting variables and the correlation between restatements and government contracting variables are negative and significant at a 1% level, consistent with our hypothesis. All significant pair-wise correlations are less than 0.50, with the exception of correlations among three size related variables (lnAsset, lnAnalyst, and lnAuditfee) and among three audit committee variables (ACSize, ACExperts, and Independence), all of which are expected. We also conduct multicollinearity tests and find that VIFs for the independent variables are well below 5; multicollinearity should not be a concern in our setting. <Insert Table 4> Electronic copy available at: https://ssrn.com/abstract=3807526 5. Empirical Results Accruals Management Table 5 report the results of estimating Equation (3) using ordinary least squares regression (OLS). The reported t-statistics are based on robust standard errors adjusting for clustering at the firm level to address heteroscedasticity. Panel A of Table 6 reports results where AbsAccurals is the dependent variable and the government contracting test variable is either GovContract (Model 1) or ContractSize (Model 2). The adjusted R for the estimation is approximately 18% for both models. The estimated coefficient on GovContract is negative and significant (-0.009, p-value < 0.01) in Model 1, which is consistent government contractors being associated with a lower level of the absolute value of discretionary accruals. The estimated coefficient on the ContractSize is negative and significant (-0.011, p-value < 0.01), consistent with the argument that the government contract values are associated with higher financial reporting quality. These results suggest that the absolute level of discretionary accruals for government contractors are 1% lower than non-government contractors. These results are economical significant given the average total asset among government contractors are $540 millions and a one percent decrease is equivalent to $5.4 million less discretionary accruals. Our hypothesis is supported. For the control variables, we observe negative and significant coefficients on lnAsset, FirmAge LnAnalyst, Independence, ACExperts, and Big4 consistent with our expectation that Average total asset for government contracotrs is 6.293 from Table 3 panel B. Exp(6.239) = 540. Electronic copy available at: https://ssrn.com/abstract=3807526 older, larger firms and those with stronger governance are associated with lower discretionary accruals. However, we observe a negative and significant coefficients on SalesVol and E/P. The coefficient on SalesVol is only marginally significant but provides some indications that higher sales volatility is associated with lower discretionary accrual during our sample period. Although we did not predict the sign for E/P, this finding is consistent with Ham et al. (2017). Moreover, the coefficients on Leverage, ACSize, and LnAudtenure are positive and significant, suggesting that high leverage firms and those with larger audit committees and longer auditor tenure are associated with higher discretionary accruals. We also observe positive and significant coefficients on, M/B, Loss, %∆CashSales, CfoVolit, and CEOChair in line with our prediction that growth firms, firms with operational issues or weaker governance are associated with higher discretionary accruals. Panel B of Table 5 reports results using AccrualQuality as the dependent variable where we vary the test variable between GovContract (Model 1) and ContractSize (Model 2). The adjusted R for the both estimations is approximately 3%. The estimated coefficient on GovContract is negative and significant (-0.002 in Model 1, p-value < 0.01), and the estimated coefficient on ContractSize is negative and significant (-0.003 in Model 2, p-value < 0.01). These results are,consistent with our hypothesis that government contractors are associated with higher accruals quality. As for our control variables, larger firms are associated with higher accruals quality. Firms that are experiencing operational difficulties and higher cashflow volatility are associated with lower accrual quality. Overall, our results are consistent with those reported by Ham et al. (2017). <Insert Table 5> Electronic copy available at: https://ssrn.com/abstract=3807526 Internal Control Weaknesses Table 6 present the results of estimating Equation (4) using Logit. The reported z- statistics are based on heteroscedasticity robust standard errors adjusting for clustering at the firm level and include industry and year fixed effects. Pseudo R for the estimations are approximately 11% for both models. The estimated coefficients on GovContract and ContractSize are negative and significant (-0.260, p-value < 0.01;-0.309, p-value < 0.01, respectively). Our results support our hypothesis and are consistent with the argument that both government contractors and the contract value are associated with a lower probability of internal control weaknesses. Firms that are larger, more mature, profitable, higher leveraged, with more AC experts, or audited by Big 4 firm are associated with lower probability of internal control material weakness. Also, higher audit fees and longer auditor tenures are associated with higher probability of internal control material weakness. <Insert Table 6> We estimate a multinomial logit regression using the number of internal control material weaknesses s as an alternative dependent variable (not tabulated). We find results consistent with those reported in Table 6. The estimated coefficients on both GovContract and ContractSize are negative and significant (-0.263, p-value < 0.01; -0.313, p-value < 0.01, respectively), The sample size is reduced to 47,185 firm-year observations as the internal control weakness data coverage start at 2004.. Electronic copy available at: https://ssrn.com/abstract=3807526 consistent with the argument that the government contractors are associated with a lower number of internal control weaknesses, supporting our hypothesis. Restatements Table 7 presents the set of tests for financial reporting quality by estimating Equation (5) for Restatements. The reported z-statistics are based on robust standard errors at the firm level to address heteroscedasticity. The pseudo R for the estimation is approximately 8% for both models. The estimated coefficients on the GovContract and ContractSize are both negative and significant (-0.199, p-value < 0.01; -0.235, p-value < 0.01, respectively). These results are consistent with the government contractors being associated with a lower probability of restatements, supporting our hypothesis. Consistent with findings on internal control material weaknesses, we find the probability of restatement is lower for firms that are larger, more mature, have higher earnings to price ratios, more analysts following, more AC experts, or are audited by a Bif 4 firm. Also, greater leverage, audit fees and percentage change in cash sales are associated with a higher probability of restatements. <Insert Table 7> Intentional mistatements are considered as another proxy for financial reporting quality. Our dependent variable in this model takes the value of one if the government contractor’s financial statements are subsequently restated because of a fraud-related class action lawsuit, and zero otherwise. Our results (untabulated) for the estimated coefficients on the GovContract and ContractSize are both negative and significant (-0.338, p-value < 0.05; -0.420 p-value < 0.05, respectively). These results are consistent with those reported in Table 7 for Restatements and Electronic copy available at: https://ssrn.com/abstract=3807526 the argument that government contractors are associated with lower probabilities of financial statement fraud, supporting our hypothesis. Additional Analyses In the following paragraphs, we discuss a number of additional analysis to further validate the association of government contracts with firm-level financial reporting quality. Change Analysis and Government Contractors We perform a analysis on the effects of change in government contract size on the change in discretionary accruals. Table 6 suggests that government contract size exhibits downward pressure on the contractor's discretionary accruals, suggesting that the association with higher firm-level financial reporting quality may be dependent on contract size. We therefore limit our sample to government contractors and use the change in government contract sales as a percentage of the contractor's total firm sales as our variable of interest. The advantage is that this approach provides more reliable and robust evidence on the effect of government contracts on firm-level financial reporting quality. We report the estimation results in Table 8 and find consistent results across the two financial reporting measures, a significant negative relationship between the change of government contract size and the change in discretionary accruals. The estimated coefficients on the ∆ContractSize are negative and significant (-0.018, p-value < 0.05; -0.009, p-value < 0.05), consistent with our hypothesis. As contract size increases, discretionary accruals are lower and accounting quality is higher. <Insert Table 8> Propensity-Score Matching Model Electronic copy available at: https://ssrn.com/abstract=3807526 In order to reduce concerns regarding misspecification of the functional form whereby the treatment of government contractors is dissimilar to the treatment of non-government contractors, we use propensity-score matching (PSM) to provide stronger controls for the effects of various firm characteristics that are separate from our variables of interest. Although Shipman, Swanquist and White (2017) advise that this method does have weaknesses, we use PSM in combination with our main model to control for the effects of observable firm characteristics and increase the validity of our primary findings. Following the PSM model suggested by Shipman et al. (2017), we estimate a Logit model explaining the probability of being a government contractor using the control variables from Equation (3). Using a caliper distance matching method with a maximum distance of one percent and without replacement, we match firms that were government contractors in any given year with firms that were never government contractors during our sample period. This procedure produces a sub-sample of firms which do not differ based on pre-existing firm characteristics but have a different treatment effect (i.e., GovContract = 1 vs. GovContract = 0). Therefore, the difference in the two subsamples is whether a firm is a government contractor or not, and provides more robust finding in testing the impact of contracting with the government on firm-level financial reporting quality. The matching procedures provide a sample that does not have significant differences in firm characteristics between firms with and without government contracts by comparing the mean and median values of two subsamples (untabulated). Electronic copy available at: https://ssrn.com/abstract=3807526 We estimate Equations (3) to (5) using the PSM sample (untabulated) and our results for the variables of interest (GovContract and ContractSize) are consistent with the results for the full sample across all models, supporting our hypothesis. Overall, these results reduce the aforementioned concerns regarding the misspecification of functional form whereby treatment firms are dissimilar to control firm. Alternative Measures for Financial Reporting Quality We consider three additional financial reporting quality measures to substantiate the robustness of our results. As discussed in the following paragraphs, our results are consistent with our earlier findings the government contracting is associated with higher financial reporting quality. First, prior research recognizes the important role of accounting conservatism for firm’s contracting (e.g., Hui, Klasa, and Yeung 2012). Goh and Li (2011) find that firms with a stronger internal control environment are more likely to understand the role of conservatism in contracting and reducing agency conflicts. Therefore, we expect federal contractors to commit to producing conservative financial statements that reduce the contracting-related transaction costs. We use the C-Score from Khan and Watts (2009) that is based on Basu’s (1997) measure of asymmetric timeliness of earnings. A higher C-Score indicates greater conservtism. We then regress C-Score on the existence of a government contracting relationship and the control variables from Equation (3) and report the results as Model (1) in Table 9. The estimated coefficient on GovContract is positive and significant (0.002, p-value < 0.05) suggesting that government contractors are associated with higher accounting conservtism as compares to non- contractors. <Insert Table 9> Electronic copy available at: https://ssrn.com/abstract=3807526 Second, earnings management can occur through accruals and/or real activities management. The difference is that accruals management misrepresents the firm’s underlying operating performance but does not involve altering firm’s operations. Real activities management involves altering normal business activities with intention to mislead its stakeholders into believing that the reported financial performance is achieved in the normal course of business (Roychowdhury 2006). Since federal government agencies are important stakeholders, investigating whether federal contractors are associated with real activities management is also necessary. We follow Cohen, Dey, and Lys (2008) and combine measures for abnormal discretionary expenses, abnormal cash flows from operations, and abnormal production costs to provide an aggregate real activities earnings management measure (REM) . Higher REM indicates an increasing level of earnings management. We regress REM on the existence of a government contracting relationship and the control variables from Equation (3) and report the results as Model (2) in Table 9. The estimated coefficient on the GovContract is negative and significant (-0.028, p-value < 0.05), suggesting that government contractors are associated with a lower level of real activities earnings management compared to non-contractors. We acknowledge that the three individual real earning management variables have different implications. As result, we also estimate the three individual real earnings management proxies. In untabulated results, we find that federal contractors are significantly associated with lower abnormal discretionary expenses and abnormal cash flow from operations. However, we did not find a significant association between the federal contractor and abnormal production cost. Electronic copy available at: https://ssrn.com/abstract=3807526 Third, Hoitash and Hoitash (2018) find that higher level of financial statement complexity (FSC) is associated with greater probability of restatement risk. Higher FSC is assumed to be attributable to managers’ intentional efforts to obfuscate financial information (e.g., Li, 2008, Lo, Ramos, and Rogo, 2017). We therefore expect government contracting is associated lower FSC. Using Hoitash and Hoitash (2018)’s FSC measure (LnARC), we regress LnARC on the existence of a government contracting relationship and the control variables from Equation (3) and report the results as Model (3) in Table 9. The estimated coefficient on the GovContract is negative and significant (-0.019, p-value < 0.05), suggesting that government contractors are associated with less complex financial statement compared to non-contractors, consistent with lower efforts to obfuscate financial statement information. Endogeneity Concerns A plausible alternative explanation for our results is that federal government agencies only award firms with higher-level financial reporting quality. As a result, high financial reporting quality among government contractors is a pre-existing characteristic of these firms before becoming a government contractor. While we cannot completely rule out this possibility, we conduct a test to mitigate this concern. We argue that government contracts help a firm to improve profitability and forecast ability and therefore, the government contractor is less likely to use discretionary accruals. In this case, we should observe increased discretionary accruals if a The FRC data is publicly available on http://www.xbrlresearch.com/ starting from 2009. Electronic copy available at: https://ssrn.com/abstract=3807526 firm loses government contracts and the opposite to occur if a firm is awarded a government contract. We utilize the PSM sample and estimate a model regressing the year-to-year change in discretionary accruals and accruals quality on indicator variables capturing whether a firm gain or loses government contracts during the previous year. The dependent variable in Model (1) of Table 10 is the change in discretionary accrual (∆AbsAccruals). The estimated coefficient on GovContractLoss is positive and significant (0.020; p-value < 0.01) indicating that when the year after losing government contracts, there is a significant increase in discretionary accruals. On the other hand, the coefficient on GovContractGain is negative and significant (-0.014; p-value < 0.10) indicating when a firm awarded a federal government contract in the prior year, there is a decrease in discretionary accruals. Similar results are obtained when the dependent variable is the change in accruals quality (Model 2). Combined, these results further confirm our previous findings that having a federal government contract is associated with the level of discretionary accruals. [Insert Table 10] In untabulated results, we also consider future weakness in internal control and future restatements. Similar to our results in Table 10, we find a significant increase (decrease) in the likelihood of internal control weakness in the future associated with losses (gains) in government contracts. Results are weaker when the dependent variable is future restatements. We find increased probability of future restatements for gains in government contracts. However, the estimated coefficient on loss in government contracts is negative but insignificant. These results further confirm our previous findings that having a federal government contract influences the probability of future weakness in internal control and restatements. Electronic copy available at: https://ssrn.com/abstract=3807526 Finally, we conduct placebo tests in a manner similar to those conducted by Bertrand and Schoar (2003) in order to address the endogeneity concern that our previous findings are not driven by firms identified as government contractors but instead is driven by a correlated omitted variable. We first exclude government contractor observations and then randomly assign non- government contractors as placebo government contractors in the same percentage as identified government contractors in the original sample. If our previous results are driven primarily by similarities between firms instead of the corporate-government contracting relationship, we should find that the coefficient on the placebo government contractor is also negative and significant. If the association is due to the firm being a government contractor, the placebo government contractor coefficient should not be significant. In untabulated results we repeat our analyses and none of the placebo government contractor coefficients are statistically significant, indicating that the government contractor status drives our results. To further ensure that the procedure is truly random, we repeat this placebo procedure 30 times with similar results. Overall, our placebo tests provide further support for the conclusion that a government contractor plays a role in determining the quality of firms’ financial statements. 6. Conclusion Cohen and Li (2019) and Samuels (2021) suggest that firms with federal government contracts are associated with higher profitability and better information environments. While there is substantial literature relating to firms’ financial reporting quality, the extant literature on investigating the reporting characteristics of government contractors is scarce. Understanding the quality of government contractors’ financial reporting is important because it helps the regulators, investors, and US citizens understand how federal government agencies are allocating Electronic copy available at: https://ssrn.com/abstract=3807526 their resources that contribute to 20% of the US GDP and, on average, 3% of the government contractors total firm sales. Our study investigates different aspects of government contractors’ financial reporting quality, including discretionary accruals, internal control weaknesses, and restatements. In the additional analysis, we further demonstrate government contractors are associated with higher level of accounting conservatism, lower level of real activities earnings management and financial statement complexity. Therefore, we document a comprehensive set of evidence that government contractors are associated with higher quality financial statements than non- government contractors across different measures of financial reporting quality. Our study contributes to the growing body of literature on the characteristics of government contractors. Our results provide important implications for the regulators, investors, and US citizens that government contractors can serve as a defensive investment mechanism that reduces the information risks relative to the non-government contractor by having higher quality financial reports. However, despite our efforts to address any potential endogeneity concerns using the placebo test and PSM matching methodology, we cannot completely rule out the possibility of the corporate government contracting relationships being randomly assigned. Electronic copy available at: https://ssrn.com/abstract=3807526 References Armstrong, C., J. Blouin, A. Jagolinzer, and D. Larcker. 2015. Corporate governance, incentives, and tax avoidance. Journal of Accounting and Economics 60 (1): 1-17. Ashbaugh-Skaife, H., W. Collins, and W. Kinney. 2007. The discovery and reporting of internal control deficiencies prior to SOX-mandated audits. Journal of Accounting and Economics 44 (1-2): 166-192. Ball, R., and L. Shivakumar. 2006. The role of accruals in asymmetrically timely gain and loss recognition. Journal of Accounting Research 44 (2): 207-242. Banerjee, S., S. Dasgupta, and Y. Kim. 2008. Buyer-supplier relationships and the stakeholder theory of capital structure. The Journal of Finance 63 (5): 2507-2552. Basu, S. 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics 24 (1): 3-37. Beneish, M. 1999. The detection of earnings manipulation. Financial Analysts Journal. 55 (5): 24-36. Bertrand, M., and A. Schoar. 2003. Managing with style: The effect of managers on firm policies. The Quarterly Journal of Economics 118 (4): 1169-1208. Callahan, C., V. Vendrzyk, and M. Butler. 2012. The impact of implied facilities cost of money subsidies on capital expenditures and the cost of debt in the defense industry. Journal of Accounting and Public Policy 31 (3): 301-319. Carcello, J., D. Hermanson, T. Neal, and R. Riley. 2002. Board characteristics and audit fees. Contemporary Accounting Research 19 (3): 365-384. Chaney, P., M. Faccio, and D. Parsley. 2011. The quality of accounting information in politically connected firms. Journal of Accounting and Economics 51(1-2): 58-76. Electronic copy available at: https://ssrn.com/abstract=3807526 Chen, H., and K. Gunny. 2014. Profitability and Cost Shifting in Government Procurement Contracts. Available at SSRN: https://ssrn.com/abstract=2482605 Cheng, A. and W. Huang, and S. Zhang. 2019. Major government customer and management earnings forecasts. Working paper, available at SSRN: https://ssrn.com/abstract=3477672. Cohen, D., and B. Li. 2016. Why do firms hold less cash? A customer base explanation. Working paper, available at SSRN: https://ssrn.com/abstract=2440794. Cohen, D., and B. Li. 2019. Customer-base concentration, investment, and profitability: The US government as a major customer. The Accounting Review 95 (1): 101-132. Cohen, D., and B. Li, N Li, and Y. Lou. 2016. Major government customers and loan contract terms. Working paper, available at SSRN: https://ssrn.com/abstract=2868761. Cohen, D., A. Dey, and T. Lys. 2008. Real and Accrual-Based Earnings Management in the Pre- and Post-Sarbanes-Oxley Periods. The Accounting Review 83(3): 757-787. Cohen, L., and A. Frazzini. 2008. Economic links and predictable returns. The Journal of Finance 63 (4): 1977-2011. Davidson, K. 2018. A big reason U.S. econoour is accelerating: Government spending WSJ.online (October 25). Available at: https://www.wsj.com/articles/government-and- military-spending-fuel-u-s-growth-1540459800/. Dechow, P., W. Ge, C. Larson, and R. Sloan. 2011. Predicting material accounting misstatements. Contemporary Accounting Research 28 (1): 17-82. Dechow, P., and We. Dichev. 2002. The quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review 77 (1): 35-59. Electronic copy available at: https://ssrn.com/abstract=3807526 DeFond, M., R. Hann, and X. Hu. 2005. Does the market value financial expertise on audit committees of boards of directors? Journal of Accounting Research 43 (2): 153-193. DeFond, M., and J. Zhang. 2014. A review of archival auditing research. Journal of Accounting and Economics 58 (2-3): 275-326. Dhaliwal, D., S. Judd, M. Serfling, and S. Shaikh. 2016. Customer concentration risk and the cost of equity capital. Journal of Accounting and Economics 61 (1): 23-48. Dhaliwal, D., V. Naiker, and F. Navissi. 2010. The association between accruals quality and the characteristics of accounting experts and mix of expertise on audit committees. Contemporary Accounting Research 27 (3): 787-827. Dou, Y., M. Khan, and Y. Zou. 2016. Labor unemployment insurance and earnings management. Journal of Accounting and Economics 61 (1): 166-184. Ellis, J., E. Fee, and S. Thomas. 2012. Proprietary costs and the disclosure of information about customers. Journal of Accounting Research 50 (3): 685-727. Federal Acquisition Regulation. 2018. 48 C.F.R. § 16. Ferris, S., R. Houston, and D. Javakhadze. 2019. It is a sweetheart of a deal: Political connections and corporate‐federal contracting. Financial Review 54 (1): 57-84. Flammer, C. 2018. Competing for government procurement contracts: The role of corporate social responsibility. Strategic Management Journal 39(5): 1299-1324. Garrett, J., R. Hoitash, and D. Prawitt. 2014. Trust and financial reporting quality. Journal of Accounting Research 52 (5): 1087-1125. Ge, W., and S. McVay. 2005. The disclosure of material weaknesses in internal control after the Sarbanes-Oxley Act. Accounting Horizons 19 (3): 137-158. Electronic copy available at: https://ssrn.com/abstract=3807526 Gosman, M., and M. Kohlbeck. 2009. Effects of the existence and identity of major customers on supplier profitability: is Wal-Mart different? Journal of Management Accounting Research 21 (1): 179-201. Goh, B., and D. Li. 2011. Internal controls and conditional conservatism. The Accounting Review 86 (3): 975-1005. Ham, C., M. Lang, N. Seybert, and S. Wang. 2017. CFO narcissism and financial reporting quality. Journal of Accounting Research 55 (5): 1089-1135. Healy, P., and J. Wahlen. 1999. A review of the earnings management literature and its implications for standard setting. Accounting Horizons 13 (4): 365-383. Hennes, K., A. Leone, and B. Miller. 2008. The importance of distinguishing errors from irregularities in restatement research: The case of restatements and CEO/CFO turnover. The Accounting Review 83 (6): 1487-1519. Hoitash, R., and U. Hoitash. 2018. Measuring accounting reporting complexity with XBRL. The Accounting Review 93 (1): 259-287. Horwitz, B., and D. Normolle. 1988. Federal agency R&D contract awards and the FASB rule for privately-funded R&D. The Accounting Review 63 (3): 414-435. Hui, K., S. Klasa, and E.Yeung. 2012. Corporate suppliers and customers and accounting conservatism. Journal of Accounting and Economics 53 (2): 115-135. Itzkowitz, J. 2013. Customers and cash: How relationships affect suppliers' cash holdings. Journal of Corporate Finance 19: 159-180. Itzkowitz, J. 2015. Buyers as stakeholders: How relationships affect suppliers' financial constraints. Journal of Corporate Finance 31: 54-66. Electronic copy available at: https://ssrn.com/abstract=3807526 Jayaraman, S. 2008. Earnings Volatility, Cash Flow Volatility, and Informed Trading. Journal of Accounting Research 46 (4): 809-851. Jensen, M., and W. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3 (4): 305-360. Jensen, M. C. 1986. Agency costs of free cash flow, corporate finance, and takeovers. The American Economic Review 76 (2): 323-329. Jones, J. 1991. Earnings management during import relief investigations. Journal of Accounting Research 29 (2): 193-228. Kalwani, M., and N. Narayandas. 1995. Long-term manufacturer-supplier relationships: do they pay off for supplier firms? Journal of Marketing 59 (1): 1-16. Karpoff, J., D. Lee, and V. Vendrzyk. 1999. Defense procurement fraud, penalties, and contractor influence. Journal of Political Economy 107 (4): 809-842. Khan, M., and R. Watts. 2009. Estimation and empirical properties of a firm-year measure of accounting conservatism. Journal of Accounting and Economics 48 (2-3): 132-150. Kinney, W, and L. McDaniel. 1989. Characteristics of firms correcting previously reported quarterly earnings. Journal of Accounting and Economics 11 (1): 71-93. Klein, A. 2002. Audit committee, board of director characteristics, and earnings management. Journal of Accounting and Economics 33 (3): 375-400. Kothari, S., A. Leone, and C. Wasley. 2005. Performance matched discretionary accrual measures. Journal of Accounting and Economics 39 (1): 163-197. Krishnan, J. 2005. Audit committee quality and internal control: An empirical analysis. The Accounting Review 80 (2): 649-675. Electronic copy available at: https://ssrn.com/abstract=3807526 Larcker, D., S. Richardson, and We. Tuna. 2007. Corporate governance, accounting outcomes, and organizational performance. The Accounting Review 82 (4): 963-1008. Li, F. 2008. Annual report readability, current earnings, and earnings persistence. Journal of Accounting and Economics 45 (2-3): 221-247. Lichtenberg, F. 1992. A perspective on accounting for defense contracts. The Accounting Review 67 (1): 741-741. Lo, K., F. Ramos. and R. Rogo. 2017. Earnings management and annual report readability. Journal of Accounting and Economics 63 (1): 1-25. McGowan, A., and V. Vendrzyk. 2002. The relation between cost-shifting and segment profitability in the defense-contracting industry. The Accounting Review 77 (4): 949-969. McNichols, M. 2002 Discussion of the quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review 77 (1): 61-69. McNichols, M. 2000. Research design issues in earnings management studies, Journal of Accounting and Public Policy 19 (4-5): 313-345. Mills, L., S. Nutter, and C. Schwab. 2012. The effect of political sensitivity and bargaining power on taxes: Evidence from federal contractors. The Accounting Review 88 (3): 977- Pownall, G. 1986. An empirical analysis of the regulation of the defense contracting industry: The Cost Accounting Standards Board. Journal of Accounting Research 24 (2): 291-315. Raman, K., and H. Shahrur. 2008. Relationship-specific investments and earnings management: Evidence on corporate suppliers and customers. The Accounting Review 83 (4): 1041- Electronic copy available at: https://ssrn.com/abstract=3807526 Reichelstein, S. 1992. Constructing incentive schemes for government contracts: An application of agency theory. The Accounting Review 67 (4): 712-731. Roychowdhury, S. 2006. Earnings management through real activities manipulation. Journal of Accounting and Economics 42 (3): 335-370. Samuels, D. (2021). Government procurement and changes in firm transparency. The Accounting Review 96 (1): 401-430. Sankar, M., and K. Subramanyam. 2001. Reporting discretion and private information communication through earnings. Journal of Accounting Research 39 (2): 365-386. Singer, Z., and J. Zhang. 2018. Auditor tenure and the timeliness of misstatement discovery. The Accounting Review 93 (2): 315-338. Shipman, J.E., Q.T. Swanquist, and R.L. White. 2016. Propensity score matching in accounting research. The Accounting Review 92 (1): 213-244. Tucker, J., and P. Zarowin. 2006. Does income smoothing improve earnings informativeness? The Accounting Review 81 (1): 251-270. Watts, R., and J. Zimmerman. 1978. Towards a positive theory of the determination of accounting standards. The Accounting Review 53 (1): 112-134. Electronic copy available at: https://ssrn.com/abstract=3807526 Appendix A Variable Definitions Corporate-government contracting variables: GovContract Indicator variable equal to one if the company is a government contractor in a given year, and zero otherwise ContractSize Contract award size relative to the firm’s total sales GovContractLost Indicator variable equal to one if the company was a government contractor in the prior year but no longer a government contractor in the current year, and zero otherwise GovContractGain Indicator variable equal to one if the company was a government contractor in the current year but not a government contractor in the prior year, and zero otherwise Financial reporting quality variables: AccrualQuality Standard deviation of residuals from t-4 to t from the Dechow and Dichev (2002) model, as modified by McNichols (2002): CA = α + α CFO + i,t 1 2 i,t-1 α CFO + α CFO + α ΔSALES + α PPE + ɛ where CA is current 3 i, t 4 i, t+1 5 i,t 6 i,t i,t, accruals, CFO is cash flow from operations, ΔSALES is change in sales and PPE is plant, property and equipment AbsAccruals Absolute value of abnormal accruals using the modified Jones (1991) approach and including the firm performance adjustment suggested by Kothari et al. (2005) ICMW Indicator variable equal to one if there is a general (systemic) material weakness in internal control reported under either SOX 302 and SOX 404 in a given year, and zero otherwise Electronic copy available at: https://ssrn.com/abstract=3807526 Restatement Indicator variable equal to one if the government contractor’s financial statement for that year is subsequently restated, and zero otherwise C_Score Conditional conservatism score developed by Khan and Watts (2009) REM Aggregate measure of real activity earnings management that equals the sum of the three standardized real earnings management metrics developed by Roychowdhury (2006) LnARC Natural logarithm of accounting reporting complexity measure developed by Hoitash and Hoitash (2018) Firm-level control variables lnAuditFees Natural logarithm of audit fees lnAsset Natural logarithm of total assets CfoVolit Standard deviation of cash flows divided by total assets over the prior five- year period SalesVol Standard deviation of sales divided by total assets over the prior five-year period E/P Current year earnings before extraordinary items divided by prior year’s market value of common equity M/B Book value of equity divided by the market value of common equity Leverage Ratio of total debt to total assets FirmAge Number of years since the firm first appeared in the CRSP database Loss Indicator variable equal to one if the firm reports losses in a given year, and zero otherwise %ΔCashSales Percent change in sales minus change in accounts receivable in the past year Electronic copy available at: https://ssrn.com/abstract=3807526 ΔROA Change in net income divided by average total assets in the past year LnAnalyst Natural logarithm of the number of analyst following the firm CEOChair Indicator variable equal to one if the CEO also serves as board chair, and zero otherwise ACSize Number of audit committee members ACExperts Proportion of financial experts on the audit committee, where financial expertise is defined as having one of the qualifications in the background of the member: certified public accountant, chief financial officer, chief accounting officer, controller, treasurer or vice-president for finance Independence Proportion of independent members on the board of directors Big4 Indicator variable equal to one if a firm’s auditor is a Big4 auditor, and zero otherwise LnAudtenure Natural logarithm of auditor’s tenure in years %Loss Percentage of loss years over the prior four years SalesGrowth Cumulative percentage change in sales over the prior three years Inventory Inventory divided by total assets ΔInventory Change in inventory divided by average total assets in the past year ΔReceivables Change in accounts receivable divided by average total assets in the past year st th Note: Continuous variables are winsorized at the1 and 99 percentile each year. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 1 Sample Selection Total number of observations in CRSP-Compustat (2001-2017), excluding firms in the banking and utility industries 115,572 Less firm-year observations: Missing data to calculate discretionary accruals 41,392 Missing data for control variables 8,712 Missing Audit Analytics data 6,496 Sample size for discretionary accruals and restatement tests 58,972 Less observations from 2001-2003 11,696 Sample size for internal control material weakness tests 47,276 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 2 Sample Distributions Panel A: Sample Distribution by Year Frequency Frequency of Gov’t Contract Gov’t Contractor % of Gov’t Year of Total % Gov’t % Value Total Sales Sales to Total Sample Contractor ($ millions) ($ millions) Sales 2001 6.28 752 4.71 75,730 1,986,205 3.8% 3,705 2002 6.73 881 5.52 97,929 2,270,550 4.3% 3,967 2003 6.82 959 6.01 114,835 3,133,326 3.7% 4,024 2004 6.72 1,060 6.64 120,039 3,633,572 3.3% 3,963 2005 6.55 1,126 7.06 113,957 4,071,181 2.8% 3,864 2006 6.35 1,138 7.13 170,576 4,623,183 3.7% 3,746 2007 6.19 1143 7.16 159,985 5,273,840 3.0% 3,650 2008 5.85 1108 6.94 199,653 5,355,060 3.7% 3,452 2009 5.63 1049 6.57 161,005 4,696,326 3.4% 3,321 2010 5.49 1018 6.38 148,280 5,070,994 2.9% 3,235 2011 5.31 967 6.06 143,837 5,624,189 2.6% 3,133 2012 5.27 906 5.68 159,579 5,101,447 3.1% 3,105 2013 5.33 847 5.31 109,682 4,922,144 2.2% 3,146 2014 5.58 807 5.06 116,697 4,851,576 2.4% 3,291 2015 5.46 774 4.85 111,875 4,865,324 2.3% 3,218 2016 5.27 775 4.86 129,098 4,571,868 2.8% 3,108 2017 5.16 679 4.25 146,906 4,733,035 3.1% 3,044 Total 58,972 100.00 15,959 100.00 2,279,663 74,783,820 3.0% Electronic copy available at: https://ssrn.com/abstract=3807526 Table 2 (continued) Panel B: Sample Distribution by Industry (Fama-French 12) % of Gov’t Gov’t Frequency Frequency % of Gov’t Contract Contractor Industry of Total of Gov’t Industry Sales to Value Total Sales Sample Contractors Total Total ($ millions) ($ millions) Sales Consumer Non-Durables 3,441 770 22.4% 16,609 5,123,382 0.3% Consumer Durables 1,920 526 27.4% 52,211 4,886,591 1.1% Manufacturing 7,383 2,592 35.1% 1,068,203 11,900,000 9.0% Energy, Oil, and Gas 3,255 214 6.6% 5,315 3,178,877 0.2% Products Chemicals and Allied 2,024 502 24.8% 8,050 2,901,472 0.3% Products Computers and Software 16,608 5,674 34.2% 712,129 15,500,000 4.6% Telephone and TV 9 - 0.0% 16,609 5,123,382 0.3% Transmission Utilities Excluded Wholesale and Retail 6,826 1,440 21.1% 140,743 17,900,000 0.8% Services Healthcare, Medical 10,603 2,627 24.8% 59,778 8,601,541 0.6% Finance Excluded Other 6,903 1,614 23.3% 216,623 4,791,957 4.5% Total 58,972 15,959 27.3% 2,279,663 74,783,820 3.0% Electronic copy available at: https://ssrn.com/abstract=3807526 Table 3 Summary Statistics Panel A: Summary Statistics for Full Sample th th Variable Standard 25 75 (N=58,972) Mean Deviation Percentile Median Percentile AbsAccruals 0.140 0.248 0.031 0.071 0.151 AccrualQuality 0.003 0.093 -0.027 0.000 0.034 IneffControls 0.046 0.209 0.000 0.000 0.000 NumWeaknesses 0.110 0.717 0.000 0.000 0.000 Restatement 0.080 0.271 0.000 0.000 0.000 Fraud 0.014 0.117 0.000 0.000 0.000 C_Score 0.127 0.139 0.054 0.142 0.206 REM -0.156 0.857 -0.443 -0.194 0.032 LnARC 5.203 0.469 4.887 5.223 5.518 CorpGov 0.271 0.444 0.000 0.000 1.000 ContractSize 0.317 0.360 0.100 0.100 0.700 lnAsset 5.520 2.321 3.857 5.522 7.174 M/B 2.514 5.226 1.158 1.631 2.599 FirmAge 10.769 8.715 5.000 9.000 14.000 Leverage 0.238 0.471 0.003 0.147 0.327 Loss 0.426 0.494 0.000 0.000 1.000 CfoVolit 0.084 0.237 0.015 0.035 0.073 SalesVol 0.208 0.479 0.042 0.109 0.219 %∆CashSales 0.183 1.027 -0.063 0.058 0.217 ∆ROA 0.005 0.228 -0.027 0.000 0.023 E/P -0.254 1.563 -0.117 0.019 0.056 CEOChair 0.308 0.461 0.000 0.000 1.000 LnAnalyst 4.516 6.469 0.000 1.778 6.417 ACSize 3.903 1.112 3.000 4.000 4.000 Independence 0.667 0.181 0.571 0.667 0.818 ACExperts 0.293 0.304 0.000 0.250 0.400 Big4 0.674 0.469 0.000 1.000 1.000 LnAudtenure 1.473 0.974 0.693 0.693 2.197 %Loss 0.399 0.384 0.000 0.250 0.750 SalesGrowth 1.988 14.333 -0.061 0.171 0.636 Inventory 0.111 0.131 0.001 0.069 0.170 lnAuditFees 13.221 1.472 12.123 13.252 14.251 ∆Inventory 0.002 0.022 -0.001 0.000 0.007 ∆Receivables 0.004 0.033 -0.005 0.002 0.013 GovContractLoss 0.029 0.168 0.000 0.000 0.000 GovContractGain 0.025 0.156 0.000 0.000 0.000 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 3 (continued) Panel B: Firms with and without Government Contracts Gov’t Contractors Non-Gov’t Contractors p-value (N=15,959) (N=43,013) for Std. Std. Mean t-test Variable Mean Dev. Median Mean Dev. Median Diff. of means AbsAccruals 0.099 0.154 0.058 0.155 0.273 0.078 -0.056 <0.001 AccrualQuality 0.000 0.072 0.000 0.004 0.100 0.000 -0.004 <0.001 IneffControls 0.046 0.209 0.000 0.046 0.209 0.000 0.000 0.963 NumWeaknesses 0.114 0.751 0.000 0.109 0.703 0.000 0.005 0.452 Restatement 0.075 0.263 0.000 0.082 0.274 0.000 -0.007 -0.005 Fraud 0.013 0.112 0.000 0.014 0.118 0.000 -0.001 -0.184 C_Score 0.097 0.14 0.113 0.141 0.137 0.156 -0.044 <0.001 REM -0.262 0.591 -0.248 -0.115 0.938 -0.171 -0.147 <0.001 LnARC 5.249 0.462 5.296 5.184 0.47 5.193 0.065 <0.001 lnAsset 6.293 2.265 6.358 5.233 2.275 5.204 1.060 <0.001 M/B 2.149 1.834 1.630 2.649 6.011 1.632 -0.500 <0.001 FirmAge 12.721 9.764 11.000 10.045 8.174 8.000 2.677 <0.001 Leverage 0.211 0.366 0.160 0.248 0.504 0.141 -0.037 -0.052 Loss 0.314 0.464 0.000 0.467 0.499 0.000 -0.153 <0.001 CfoVolit 0.059 0.154 0.031 0.094 0.260 0.037 -0.034 <0.001 SalesVol 0.187 0.343 0.115 0.216 0.521 0.105 -0.029 <0.001 %∆CashSales 0.123 0.551 0.065 0.205 1.154 0.054 -0.081 <0.001 ∆ROA 0.002 0.131 0.000 0.006 0.255 -0.001 -0.004 0.065 E/P -0.128 1.244 0.033 -0.300 1.664 0.009 0.172 <0.001 CEOChair 0.400 0.490 0.000 0.273 0.446 0.000 0.127 <0.001 LnAnalyst 6.255 7.233 3.917 3.871 6.036 1.000 2.383 <0.001 ACSize 4.041 1.150 4.000 3.851 1.093 4.000 0.190 <0.001 Independence 0.665 0.163 0.667 0.668 0.187 0.667 -0.003 <0.001 ACExperts 0.370 0.314 0.333 0.265 0.296 0.250 0.105 <0.001 Big4 0.761 0.426 1.000 0.641 0.480 1.000 0.120 <0.001 LnAudtenure 1.671 1.101 1.386 1.400 0.911 0.693 0.272 <0.001 %Loss 0.303 0.358 0.250 0.434 0.387 0.500 -0.131 <0.001 SalesGrowth 0.925 7.913 0.204 2.382 16.057 0.153 -1.458 <0.001 Inventory 0.125 0.124 0.097 0.106 0.133 0.055 0.019 <0.001 lnAuditFees 13.791 1.466 13.875 13.009 1.417 13.016 0.782 <0.001 ∆Inventory 0.003 0.021 0.000 0.002 0.023 0.000 0.001 <0.001 ∆Receivables 0.004 0.028 0.003 0.004 0.034 0.002 0.001 0.015 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 3 (continued) st th All variables are defined in Appendix A. Continuous variables are winsorized at the1 and 99 percentile each year. N= 47,276 overall, N=13,337 for Government contractors, and N=33,871 for non-Government contractors. Internal control weakness data coverage start in 2004. Therefore, there are fewer observations for internal control weakness tests. N=44,410 for C-Score (N=13,708 for Government contractors and N=30,702 for non- Government contractors), 54,420 for REM (N=15,364 for Government contractors and N=39,056 for non-Government contractors), and 18,698 for LnARC measures (N=5,352 for Government contractors and N=13,346 for non-Government contractors). LnARC is available for substanial fewer observations because the financial complexity data coverage starts in 2009. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 4 Pairwise Correlations Panel A: Variables 1 to 10 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) 1.00 (2) 0.03* 1.00 (3) -0.01* -0.02* 1.00 (4) 0.05* -0.02* 0.21* 1.00 (5) -0.10* -0.02* 0.00 -0.01* 1.00 (6) -0.10* -0.02* 0.00 -0.01* 0.99* 1.00 (7) -0.32* -0.05* 0.04* -0.01 0.20* 0.20* 1.00 (8) 0.23* 0.02* -0.01* 0.00 -0.04* -0.04* -0.18* 1.00 (9) -0.06* -0.01 -0.03* -0.01 0.14* 0.14* 0.15* -0.05* 1.00 (10) 0.16* 0.00 0.00 0.02* -0.04* -0.03* -0.05* 0.22* 0.02* 1.00 (11) 0.24* 0.00 0.02* 0.02* -0.14* -0.13* -0.42* 0.11* -0.15* 0.10* (12) 0.17* 0.04* -0.02* 0.00 -0.06* -0.06* -0.31* 0.16* -0.01 0.20* (13) 0.05* 0.01 -0.01 0.00 -0.03* -0.03* -0.20* 0.01* 0.02* 0.06* (14) 0.10* 0.00 0.00 0.01* -0.04* -0.03* -0.04* 0.03* -0.06* -0.01 (15) 0.03* 0.15* -0.01* -0.01 -0.01 -0.01 -0.03* -0.02* -0.02* -0.04* (16) -0.10* 0.06* -0.01 -0.01* 0.05* 0.05* 0.11* 0.01 0.03* -0.17* (17) -0.10* 0.00 0.01 0.01 0.12* 0.12* 0.22* -0.03* 0.08* -0.04* (18) -0.21* -0.04* 0.00 -0.03* 0.19* 0.19* 0.64* -0.04* 0.05* -0.06* (19) -0.17* -0.02* 0.03* 0.01 0.20* 0.20* 0.41* -0.07* 0.16* -0.09* (20) -0.14* -0.01* 0.00 0.00 0.15* 0.14* 0.33* -0.05* 0.10* -0.06* (21) -0.15* -0.01* 0.02* 0.00 0.17* 0.16* 0.28* -0.05* 0.09* -0.11* (22) -0.11* -0.01* 0.00 -0.03* 0.05* 0.05* 0.26* -0.05* -0.03* -0.04* (23) -0.14* -0.02* 0.00 0.00 0.14* 0.13* 0.32* -0.06* 0.21* -0.03* (24) 0.26* 0.02* 0.01* 0.02* -0.15* -0.15* -0.50* 0.13* -0.15* 0.11* (25) 0.06* 0.01 0.00 0.02* -0.05* -0.04* -0.04* 0.02* -0.06* 0.01 (26) -0.07* 0.02* 0.00 0.00 0.07* 0.06* -0.04* -0.07* 0.12* 0.01* (27) -0.24* -0.05* 0.10* 0.02* 0.24* 0.23* 0.86* -0.12* 0.15* -0.02* (28) 0.03* 0.24* -0.01 0.01 0.02* 0.01* 0.06* 0.02* -0.01* -0.03* (29) 0.06* 0.26* 0.00 0.00 0.01 0.01 0.05* 0.03* -0.03* -0.04* Electronic copy available at: https://ssrn.com/abstract=3807526 Table 4 (continued) Panel B: Variables 11 to 20 (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (11) 1.00 (12) 0.20* 1.00 (13) 0.09* 0.46* 1.00 (14) 0.08* -0.03* -0.07* 1.00 (15) 0.13* -0.04* -0.03* 0.08* 1.00 (16) -0.17* -0.11* -0.10* 0.03* 0.11* 1.00 (17) -0.16* -0.08* -0.05* -0.01 -0.01* 0.07* 1.00 (18) -0.30* -0.18* -0.15* -0.01* -0.02* 0.13* 0.22* 1.00 (19) -0.19* -0.14* -0.10* -0.02* -0.02* 0.12* 0.43* 0.41* 1.00 (20) -0.15* -0.11* -0.09* -0.02* -0.01* 0.10* 0.28* 0.35* 0.60* 1.00 (21) -0.11* -0.12* -0.11* -0.01 -0.01* 0.12* 0.40* 0.33* 0.85* 0.66* (22) -0.09* -0.08* -0.05* -0.01 -0.01* 0.02* 0.06* 0.21* 0.10* 0.09* (23) -0.17* -0.07* -0.04* -0.05* -0.02* 0.06* 0.11* 0.25* 0.24* 0.17* (24) 0.77* 0.27* 0.13* 0.08* 0.08* -0.16* -0.17* -0.35* -0.21* -0.16* (25) 0.06* -0.02* -0.04* 0.15* 0.03* -0.02* -0.02* -0.03* -0.04* -0.04* (26) -0.13* -0.03* 0.06* -0.05* -0.02* 0.00 0.01* -0.08* 0.01* 0.01 (27) -0.30* -0.20* -0.14* -0.04* -0.03* 0.08* 0.21* 0.59* 0.45* 0.39* (28) -0.12* -0.09* -0.11* 0.13* 0.04* 0.11* 0.04* 0.05* 0.04* 0.04* (29) -0.06* -0.11* -0.13* 0.15* 0.09* 0.13* 0.04* 0.05* 0.03* 0.03* Panel C: Variables 21 to 29 (21) (22) (23) (24) (25) (26) (27) (28) (29) (21) 1.00 (22) 0.08* 1.00 (23) 0.17* 0.09* 1.00 (24) -0.11* -0.10* -0.20* 1.00 (25) -0.03* 0.00 -0.06* 0.08* 1.00 (26) -0.01* -0.03* 0.04* -0.15* -0.04* 1.00 (27) 0.36* 0.25* 0.30* -0.36* -0.04* -0.05* 1.00 (28) 0.05* 0.01 -0.01* -0.08* 0.04* 0.16* 0.03* 1.00 (29) 0.04* 0.00 -0.02* -0.04* 0.05* -0.03* 0.02* 0.27* 1.00 Table 4 shows Pearson correlation coefficients for variables included in Equations: (1) AbsAccruals (2) AccrualQuality (3) ICMW (4) Restatement (5) CorpGov (6) ContractSize (7) lnAsset (8) M/B (9) FirmAge (10) Leverage (11) Loss (12) CfoVolit (13) SalesVol (14) %∆CashSales (15) ∆ROA (16) E/P (17) CEOChair (18) LnAnalyst (19) ACSize (20) ACExperts (21) Independence (22) Big4 (23) LnAudtenure (24) %Loss (25)SalesGrowth (26)Inventory (27) lnAuditFees (28) ∆Inventory (29) ∆Receivables. All variables are defined in the Appendix. * represent significance at the 0.01 level. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 5 Regression of Accruals Management on Government Contracting AbsAccruals (or AccrualQuality) = β + β CORP_GOV + β lnAsset + β M/B + 0 1 2 3 β Leverage + β FirmAge + β Loss + β %ΔCashSales + β ΔROA + β CfoVolit + 4 5 6 7 8 9 β SalesVol + β E/P+ β lnAnalyst + β CEOChair + β ACSize + 10 11 12 13 14 β Independence + β ACExperts + β Big4 + β LnAudtenure + Fixed Effects + ɛ 15 16 17 18 Panel A: Dependent Variable Equals Absolute Value of Abnormal Accruals Model (1) Model (2) Variables Pred Coefficients t-stat. Coefficients t-stat. *** *** Constant 0.213 12.54 0.214 12.62 *** GovContract H1: - -0.009 4.14 *** ContractSize H1: - -0.011 4.09 *** *** lnAsset - -0.020 19.59 -0.020 19.61 *** *** M/B + 0.006 4.94 0.006 4.94 *** *** Leverage ? 0.045 5.09 0.045 5.10 *** *** FirmAge - -0.004 2.91 -0.004 2.91 *** *** Loss + 0.038 13.92 0.038 13.93 *** *** %∆CashSales + 0.017 8.34 0.017 8.34 ∆ROA + 0.020 0.92 0.020 0.92 *** *** CfoVolit + 0.045 3.16 0.045 3.16 * * SalesVol + -0.008 1.85 -0.008 1.85 *** *** E/P ? -0.008 6.65 -0.008 6.64 * * LnAnalyst - -0.002 1.68 -0.002 1.69 *** *** CEOChair + -0.006 3.18 -0.006 3.18 *** *** ACSize ? 0.004 3.53 0.004 3.53 *** *** Independence - -0.048 5.43 -0.048 5.43 *** *** ACExperts - -0.036 10.31 -0.036 10.32 *** *** Big4 - -0.026 8.63 -0.026 8.63 ** ** LnAudtenure ? 0.002 2.18 0.002 2.18 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 18.46% 18.45% Observations 58,972 58,972 Electronic copy available at: https://ssrn.com/abstract=3807526 Table 5 (continued) Panel B: Dependent Variable Equals Accruals Quality Model (1) Model (2) Variables Pred. Coefficients t-stat. Coefficients t-stat. *** *** Constant 0.019 3.97 0.020 4.05 *** GovContract H1: - -0.002 3.04 *** ContractSize H1: - -0.003 2.91 *** *** lnAsset - -0.002 5.31 -0.002 5.33 M/B + 0.000 0.57 0.000 0.57 Leverage ? 0.001 0.64 0.001 0.65 *** *** FirmAge - -0.001 2.65 -0.001 2.65 *** *** Loss + -0.009 9.79 -0.009 9.79 ** ** %∆CashSales + -0.001 2.14 -0.001 2.14 *** *** ∆ROA + 0.064 12.01 0.064 12.01 *** *** CfoVolit + 0.019 3.42 0.019 3.42 SalesVol + -0.002 1.20 -0.002 1.20 *** *** E/P ? 0.003 6.20 0.003 6.20 *** *** LnAnalyst - -0.002 3.91 -0.002 3.92 ** ** CEOChair + 0.001 2.09 0.001 2.09 ACSize ? 0.001 1.46 0.001 1.47 Independence - -0.003 1.24 -0.003 1.24 ACExperts - 0.001 0.84 0.001 0.83 Big4 - -0.001 0.46 -0.001 0.46 LnAudtenure ? -0.001 1.06 -0.001 1.06 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 3.24% 3.25% Observations 58,972 58,972 Table 5 reports results from estimating Equation (3) using OLS. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 6 Regression of Internal Control Quality on Government Contracting Prob ( ICMW = 1) = F [ δ + δ CORP_GOV + δ lnAsset + δ M/B +δ Leverage + 0 1 2 3 4 δ FirmAge + δ %Loss + δ SalesGrowth + δ Inventories + δ LnAuditFee+ 5 6 7 8 9 δ LnAnalyst + δ CEOChair + δ ACSize + δ Independence + δ ACExperts + 10 11 12 13 14 δ Big4 + δ LnAudtenure + Fixed effects + ɛ ] 15 16 Model (1) Model (2) Variables Pred. Coefficients z-stat. Coefficients z-stat. *** *** Constant 16.842 17.55 16.809 17.51 *** GovContract H1: - -0.260 3.65 *** ContractSize H1: - -0.309 3.48 *** *** lnAsset - -0.502 11.79 -0.503 11.81 M/B - -0.026 1.21 -0.026 1.21 * * Leverage + -0.168 1.88 -0.167 1.88 *** *** FirmAge - -0.234 5.98 -0.235 5.98 ** ** %Loss + 0.204 2.17 0.204 2.17 SalesGrowth + 0.002 1.25 0.002 1.27 Inventory + 0.121 0.37 0.122 0.38 *** *** lnAuditFees + 1.288 20.27 1.288 20.28 LnAnalyst - -0.057 1.54 -0.058 1.56 CEOChair + 0.047 0.79 0.048 0.80 * * ACSize ? -0.050 1.77 -0.049 1.76 *** *** Independence - -0.349 3.26 -0.350 3.27 ACExperts - 0.063 0.32 0.064 0.33 *** *** Big4 - -0.665 7.56 -0.665 7.57 *** *** LnAudtenure ? 0.285 10.03 0.285 10.02 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Pseudo R 11.20% 11.10% Observations 47,267 47,267 Table 6 reports results from estimating Equation (4) using Logit. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 7 Regression of Accounting Restatements on Government Contracting Prob (Restatement =1) = F [δ + δ CORP_GOV + δ lnAsset + δ M/B + δ Leverage + 0 1 2 3 4 δ FirmAge + δ ΔInventories + δ ΔReceivables + δ LnAuditFee + δ %ΔCashSales + 5 6 7 8 9 δ E/P δ ΔROA + δ LnAnalyst + δ CEOChair + δ ACSize + δ Independence + 10 + 11 12 13 14 15 δ ACExperts + δ Big4 + δ LnAudtenure + Fixed effects + ɛ ] 16 17 18 Model (1) Model (2) Variables Pred. Coefficients z-stat. Coefficients z-stat. *** *** Constant 6.013 11.66 5.986 11.61 *** GovContract H1: - -0.199 4.69 *** ContractSize H1: - -0.235 4.48 *** *** lnAsset - -0.142 7.41 -0.143 7.44 M/B - -0.001 0.16 -0.001 0.17 ** ** Leverage + 0.055 2.09 0.055 2.10 *** *** FirmAge - -0.185 7.55 -0.185 7.55 ∆Inventory + 1.079 1.40 1.078 1.40 ∆Receivables + 0.392 0.75 0.394 0.75 *** *** lnAuditFees + 0.275 8.85 0.276 8.85 *** *** %∆CashSales + 0.046 3.27 0.046 3.28 *** *** E/P + -0.024 3.16 -0.024 3.15 ∆ROA - -0.124 1.57 -0.124 1.57 *** *** LnAnalyst - -0.107 4.84 -0.108 4.86 CEOChair ? -0.023 0.62 -0.023 0.61 ACSize ? 0.006 0.38 0.006 0.38 Independence ? -0.104 1.04 -0.104 1.04 ** ** ACExperts - -0.153 2.40 -0.154 2.42 *** *** Big4 - -0.447 9.39 -0.447 9.39 *** *** LnAudtenure ? 0.621 34.63 0.621 34.61 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Pseudo R 7.92% 7.91% Observations 58,972 58,972 Table 7 reports results from estimating Equation (5) using Logit. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 8 Regression of Change in Discretionary Accruals on Change in Government Contract Size ∆AbsAccruals ∆AccrualQuality Variables Coefficients t-stat. Coefficients t-stat. Constant -0.041 1.91 -0.013 1.15 ** ** ∆ContractSize -0.018 2.17 -0.009 2.04 *** ∆lnAsset 0.094 5.13 -0.006 0.67 ∆M/B 0.002 0.60 0.000 0.07 ∆Leverage 0.000 0.21 -0.000 1.28 ** FirmAge -0.092 2.52 0.022 0.90 Loss 0.009 1.14 0.000 0.05 %∆CashSales -0.053 1.11 0.060 1.40 ∆ROA 0.009 0.50 -0.002 0.18 *** ∆CfoVolit 0.004 0.59 -0.010 3.51 ∆SalesVol 0.037 0.48 0.026 1.18 ** ∆E/P 0.006 2.24 0.000 0.34 ∆LnAnalyst 0.000 0.03 0.005 1.81 *** CEOChair -0.002 3.93 -0.000 0.07 ACSize 0.002 0.87 0.000 0.10 Independence 0.002 0.12 0.007 0.69 ACExperts -0.001 0.17 -0.001 0.13 Big4 0.004 0.35 -0.005 0.84 LnAudtenure 0.001 0.13 -0.001 0.36 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 14.60% 8.86% Observations 14,576 14,576 Table 8 report results from estimating a change specification of Equation (3) using only government contractors to provide evidence on the importance of government contract size. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 9 Alternative Financial Reporting Quality Measures Model (1) Model (2) Model (3) C_Score REM LnARC Variable Coefficients t-stat. Coefficients t-stat. Coefficients t-stat. *** *** *** Constant 0.420 38.86 0.347 5.24 3.407 48.68 ** ** ** GovContract 0.002 2.00 -0.028 2.47 -0.019 2.10 *** *** *** lnAsset -0.044 105.95 -0.084 0.162 44.60 16.92 *** M/B 0.047 26.86 -0.004 1.61 -0.005 1.33 *** *** *** Leverage 0.068 17.12 0.156 3.97 0.132 8.30 *** *** FirmAge 0.000 0.51 -0.019 2.71 -0.028 5.46 *** *** *** ∆Inventory 0.010 7.83 0.229 0.056 7.03 19.27 *** *** ∆Receivables -0.003 5.74 0.050 4.95 0.001 0.36 ** ** lnAuditFees -0.016 2.55 0.207 2.28 0.005 0.52 *** *** *** %∆CashSales -0.045 6.87 0.348 5.45 0.063 2.84 *** *** E/P 0.008 3.89 0.026 1.15 0.058 4.37 * *** ∆ROA -0.005 1.68 -0.000 0.00 -0.026 5.31 *** *** LnAnalyst -0.011 17.79 0.006 0.93 -0.059 10.00 *** * *** CEOChair 0.007 6.86 0.020 1.83 -0.037 4.53 *** * ACSize -0.000 0.45 0.018 3.47 0.007 1.89 *** *** Independence -0.006 3.39 -0.002 0.12 -0.137 9.17 *** ACExperts -0.015 5.19 -0.056 1.52 0.005 0.17 *** ** Big4 -0.014 10.79 0.016 1.07 -0.025 2.06 *** ** *** LnAudtenure 0.001 2.60 0.011 1.96 0.016 3.70 Year Fixed Effects Yes Yes Yes Industry Fixed Effects Yes Yes Yes Adjusted R 0.644 0.144 0.599 Observations 44,409 54,134 18,590 Table 9 reports results from estimating Equation (3) using alternative financial reporting quality measures and OLS. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526 Table 10 Analysis of Government Contract Gains and Losses Model (1) Model (2) ∆AbsAccruals ∆AccrualQuality Variables Coefficients t-stat. Coefficients t-stat. *** *** Constant -0.025 4.20 -0.012 3.70 *** * GovContractLoss 0.020 2.90 0.007 1.67 t-1 * ** GovContractGain -0.014 1.73 -0.008 2.25 t-1 *** ∆lnAsset 0.038 7.76 -0.000 0.15 * * ∆M/B 0.004 1.92 0.002 1.82 *** ∆Leverage 0.000 3.53 0.000 0.99 *** FirmAge -0.109 3.18 0.012 0.96 *** *** Loss 0.037 9.24 0.007 3.77 *** *** %∆CashSales -0.141 3.37 0.061 3.78 ∆ROA -0.016 1.54 -0.003 0.47 ** ∆CfoVolit 0.001 0.19 -0.005 2.57 *** *** ∆SalesVol -0.609 9.78 0.146 5.94 ∆E/P 0.002 0.86 0.002 1.14 ∆LnAnalyst 0.001 0.31 -0.000 0.47 *** ** CEOChair -0.003 4.90 0.001 2.55 ACSize 0.000 0.17 -0.001 0.78 ACExperts -0.003 0.30 -0.005 1.10 Independence 0.002 0.10 0.001 0.15 ** Big4 0.003 0.91 0.004 2.55 *** LnAudtenure 0.003 3.46 -0.001 0.07 Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Adjusted R 18.50% 5.40% Observations 29,668 29,668 Table 10 report results from estimating a change specification of Equation (3) using the PSM sample and controlling for gains and losses in government contracts. All variables are defined in Appendix A. Industry fixed effects are based on SIC 2-digit codes, and reported significance is based on robust standard errors and two-tailed tests, adjusted for heteroscedasticity and clustered by firm. ***, ** and * represent significance at 1%, 5% and 10% levels, respectively. Electronic copy available at: https://ssrn.com/abstract=3807526

Journal

ARN Conferences & Meetings – SSRN

Published: Mar 18, 2021

Keywords: public procurement, government contractors, financial reporting quality

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Federal Government Contracts and Financial Reporting Quality

Federal Government Contracts and Financial Reporting Quality

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Federal Government Contracts and Financial Reporting Quality

Federal Government Contracts and Financial Reporting Quality

References (19)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies