Addressing unobserved endogeneity bias in accounting studies: control and sensitivity methods by variable type

Together with their associated statistical routines, this paper describes the control and sensitivity methods that can be employed by accounting researchers to address the important issue of unobserved (omitted) variable bias in regression and matching models according to the types of variables employed. As with other social science disciplines, an important and pervasive issue in observational (non-experimental) accounting research is omitted variable bias (endogeneity). Causal inferences for endogenous explanatory variables are biased. This occurs in regression models where an unobserved (confounding) variable is correlated with both the dependent (outcome) variable in a regression model and the causal explanatory (often a selection) variable of interest. The Heckman treatment effect model has been widely employed to control for hidden bias for continuous outcomes and endogenous binary selection variables. However, in accounting studies, limited (categorical) dependent variables are a common feature and endogenous explanatory variables may be other than binary in nature. The purpose of this paper is to provide an overview of contemporary control methods, together with the statistical routines to implement them, which extend the Heckman approach to binary, multinomial, ordinal, count and percentile outcomes and to where endogenous variables take various forms. These contemporary methods aim to improve causal estimates by controlling for hidden bias, though at the price of increased complexity. A simpler approach is to conduct sensitivity analysis. This paper also presents a synopsis of a number of sensitivity techniques and their associated statistical routines which accounting researchers can employ routinely to appraise the vulnerability of causal effects to potential (simulated) unobserved bias when estimated with conventional regression and propensity score matching estimators.     

Assessing self-selection and endogeneity issues in the relation between activity-based costing and performance

Studies investigating the relation between ABC adoption and performance are inconclusive and plagued with econometric problems. This study extends prior research to investigate the association between ABC adoption and four manufacturing plant performance measures (cycle-time improvement, quality improvement, cost improvement, and profitability) and to assess selection bias and the endogenous nature of their relationship. I use the Heckman (1979) model to assess sample selection bias and the Wooldridge (2002) 2SLS-IV approach, to investigate endogeneity. After controlling for sample selection bias and endogeneity, the coefficient of ABC under the Heckman method and ABC_fit under the 2SLS-IV method becomes significantly higher compared to the coefficient of ABC under the OLS method. In addition, both the inverse Mills ratio, under the Heckman model, and Hausman F-test, under the Wooldridge 2SLS approach, are positive and significant, confirming the presence of both sample selection bias and endogeneity. Overall, I find that controlling for sample selection bias and endogeneity is essential in properly assessing the significance of ABC-performance association.     

Lock‐In Effects in Relationship Lending: Evidence from DIP Loans

Do prior lending relationships result in pass‐through savings (lower interest rates) for borrowers, or do they lock in higher costs for borrowers? Theoretical models suggest that when borrowers experience greater information asymmetry, higher switching costs, and limited access to capital markets, they become locked into higher costs from their existing lenders. Firms in Chapter 11 seeking debtor‐in‐possession (DIP) financing often fit this profile. We investigate the presence of lock‐in effects using a sample of 348 DIP loans. We account for endogeneity using the instrument variable (IV) approach and the Heckman selection model and find consistent evidence that prior lending relationship is associated with higher interest costs and the effect is more severe for stronger existing relationships. Our study provides direct evidence that prior lending relationships do create a lock‐in effect under certain circumstances, such as DIP financing.     

The impact of replacement cost disclosure on security prices: New Evidence

A number of studies have tested for information content in the ASR 190 disclosure by comparing the conditional and unconditional distribution of abnormal security returns around the time of disclosure. Since no differences were observed, it was concluded that ASR 190 had no information content. The study reported below performs a similar test by estimating the regression function of the conditional distribution of abnormal returns. This test procedure controls for the information content in contemporaneous historical cost disclosure and uses a conditioning variable not considered in earlier tests. It finds statistically significant stock price effects. However, because most of the effects appear to precede the official announcement date by several months, it is unclear whether stock prices were responding to the leakage of the information content of ASR 190 prior to disclosure, to private production of information contained in ASR 190 or to a variable omitted from the study which happens to be correlated with replacement costs.     

Value Creation and Governance Structure in Reit Mergers

The Umbrella Partnership REIT (UPREIT) structure has become the dominant form of organization for U.S. REITs. We examine the utility of this corporate structure from a new perspective, finding evidence that convertible securities issued by UPREITs in payment for properties acquired from private sellers often function as instruments of corporate control, aligning the interests of new executives acquired in the transaction with those of the purchasing REIT’s shareholders. We also find evidence that these financial arrangements are used to signal information regarding the firm’s future prospects. We use a sample of 53 public–private mergers 1995–2001, in which the acquirer is a publicly traded REIT. We find that wealth effects from central managerial changes are positively related to the degree to which payment takes the form of convertible equity units of UPREIT subsidiaries, and to the minimum lock-up period for those units prior to conversion. The positive effects of longer lock-ups are evidence that financing structure can be used to reduce agency and information costs related to managerial restructuring in public–private mergers.     

Multiple Large Shareholders and Financial Reporting Quality: Evidence from China

In this paper, we examine the effect of multiple large shareholders (MLS) on financial reporting quality. Using a sample of Chinese listed firms over the period 2007–2018, we find that firms with MLS tend to have lower financial reporting quality. Our findings are robust to an array of robustness tests, including controlling for possible omitted variables, a Heckman two-step sample selection model, and a difference-in-differences analysis based on a propensity score matched sample. We further show that the effect of MLS on financial reporting quality is attenuated for firms followed by more analysts, cross-listed on the Hong Kong Stock Exchange, and held by institutional blockholders. Finally, we find that agency problems appear to be the possible underlying mechanisms through which MLS lower financial reporting quality.     

审计师选择与设立审计委员会的自选择问题——来自中国证券市场的经验证据

本文采用Heckman(1978)的二阶段回归方法,以2002-2006年我国证券市场上市公司为研究样本,在研究审计师选择的同时控制了设立审计委员会的自选择问题,并将公司设立审计委员会的实际情况与假设的相反情况进行对比,分析它们不同的审计师选择倾向,从而间接检验了审计委员会的治理效率。研究发现:自选择问题对审计师选择存在显著的影响。因此,拒绝了公司随机设立审计委员会的原假设;在控制自选择之后,回归方程中变量的斜率系数整体上存在显著差异;并且没有设立审计委员会的公司比设立审计委员会的公司更可能聘请四大会计师事务所。     

Directors' and officers' liability insurance and stock price crash risk

We investigate the impact of directors' and officers' insurance (D&O insurance) on stock price crash risk. We find that D&O insurance in China is negatively associated with stock price crash risk. This association is robust to a series of robustness checks including the use of alternative sample, Heckman two-step sample selection model, propensity score matching procedure, fixed effects model, the inclusion of some possibly omitted variables, and bootstrap method. Further analyses show that the impact of D&O insurance on crash risk is more pronounced in firms with lower board independence, non-Big 4 auditors, lower institutional shareholdings, and weaker investor protection; and the negative relationship between D&O insurance and crash risk is not driven by the eyeball effect. Moreover, we find that D&O insurance purchase is associated with less financial restatements and more disclosure of corporate social responsibility reports. Our findings provide support to the notion that D&O insurance appears to improve corporate governance.     

Owner liability and financial reporting information as predictors of firm default in bank loans

We examine the effects of owner liability and non-accounting and financial accounting information on the probability of default as defined in Basel II in bank loan contracted by non listed firms. We model default as a function of owner liability and accounting and non-accounting information of non-listed firms, drawing on 43,117 annual accounts of 16,029 firms over a 7-year period. Our estimations based on mixed logistic regressions with random parameters show that the predicted default probability of full-liability firms is 0.72 times that of limited liability firms. The likelihood ratio test for omitted variables confirms the additional predictive ability of liability status over and above other non-accounting and financial accounting information. A Heckman self-selection model does not indicate sampling bias. The particular definition of default used in the study enables the findings to be generalizable across other institutional contexts.     

The Role of Information Precision in Determining the Cost of Equity Capital

We examine the association between the cost of equity capital and the quality of public and private information. We find an inverse relationship between the cost of capital and the precision of public information, but the effect is more than offset by a positive relationship between the cost of equity capital and the precision of private information. Public and private information precisions are positively correlated, and a model that fails to include both is vulnerable to a correlated omitted variable bias. The association between public and private information combined with their opposing effects on the cost of capital implies mangers should consider the relationship between public and private information when assessing their reporting strategy.     

The risk and return of private equity real estate funds

This study employs an asset pricing approach to quantify the exposure of private real estate funds to both private and publicly traded real estate risk factors. The analysis includes the creation of specific performance indices and the use of methods seeking to address some of the inherent issues with private real estate fund data, such as the high degree of serial correlation in observed total returns. The use of the Driessen et al. (2012) estimator leads to markedly higher risk factor coefficient estimates than seen in the prior literature and estimated with time series regression.     

DOES IT PAY TO DISCLOSE MANAGERIAL EARNINGS INFORMATION EARLY?

Does it pay to voluntarily disclose the manager's private information about the firm's earnings prospects before the mandatory announcement date? This question has been a subject of much debate because prior research establishes both benefits and costs of early information disclosure. We provide evidence on the net effect of such disclosure by examining its impact on firm value. Using a large sample and correcting for self‐selection bias, we find that early disclosure of the manager's private earnings information enhances the end‐of‐period value of the firm.     

Geographic earnings disclosure and trading volume

Beginning with Statement of Financial Accounting Standards No. 131 (SFAS 131), Disclosures about Segments of an Enterprise and Related Information, most US multinational firms no longer disclose geographic earnings in their annual reports. Given the recent growth in foreign operations of US firms and the varying operating environments around the world, information (or lack thereof) related to geographical performance can affect investors’ information set. Using empirical tests that closely follow the [Kim, O., Verrecchia, R., 1997. Pre-announcement and event-period private information. Journal of Accounting and Economics 24, 395–419] model, we find results consistent with their predictions. Specifically, using a sample of firms with substantial foreign operations, we find evidence of a decrease in event period private information following adoption of SFAS 131 for firms that no longer disclose geographic earnings. These results suggest that decreased public information (i.e., non-disclosure of geographic earnings) reduces the ability of investors to utilize or generate private information in conjunction with the public announcement of quarterly earnings, which dampens trading. We also find evidence of a decrease in pre-announcement private information following adoption of SFAS 131. This is consistent with an overall improvement in public disclosures that has the effect of reducing differences in the precision of private information across investors in the period prior to the earnings announcement. However, such an effect is observed for both firms which no longer disclose geographic earnings and for firms that continue to disclose geographic earnings.     

Detecting the presence of insider trading via structural break tests

The occurrence of abnormal returns before the unscheduled announcement of price sensitive information is a potential indicator of insider trading. We identify insider trading with a structural change in the intercept of an extended capital asset pricing model. To detect such a change we introduce a consistent timing structural break test (CTSB) based upon a U-statistic type process. Unlike the traditional CUSUM test, the CTSB test provides a consistent estimator of the timing of a break in the intercept that occurs across the whole evaluation period. We apply our test to a rich data set covering 370 price sensitive announcements relating to FTSE 350 companies. Our test is able to detect potential insider trading far more reliably than the standard CUSUM test. We also show that the majority of suspected insider trading takes place in the 25 days prior to the release of market sensitive information.     

The Construction of Residential Housing Price Indices: A Comparison of Repeat-Sales,Hedonic-Regression,and Hybrid Approaches

This article examines a number of hypotheses that underpin the repeat-sales and hedonic approaches to the construction of housing price indices, as well as the practical problems associated with the implementation of either approach. We also examine a hybrid procedure that combines elements of both the repeat-sales and hedonic-regression techniques. For our sample of individual home sales in Oakland and Fremont California over an 18-year period, repeat-sales methods are subject to sample selection bias; the maintained assumption of time constancy of implicit prices of housing attributes is violated; the repeat-sales estimator is extremely sensitive to influential observations; and the usual method used to correct for heteroskedasticity in repeat-sale housing returns is inappropriate in our sample. Hedonic techniques are better suited to contend with index number problems per se, as they can accommodate changing attribute prices over time. They also appear to give rise to more reliable estimates of price indices, as unusual observations have less effect on estimated price indices. Drawbacks of the hedonic approach include the usual concern with omitted attributes, and their effect on the estimated price index.     

Call it good,bad or no news? The valuation effect of debt issues

Studies investigating market reactions to changes in capital structure aim to find the impact of private information conveyed. However, these studies ignore that financial decisions are not made randomly but are conditional on managers’ private information. Using a sample of U.S. leverage‐increasing public companies with public long‐term debt offerings, we find that debt offerings convey no new information to the markets after considering the conditionality of decisions. We also show that results can be biased if the deterministic role played by private information ex ante is omitted, which may explain the conflicting valuation evidence found in the literature.     

Industry and time specific deviations from fundamental values in a random coefficient model

The paper analyzes the relationship between stock prices and fundamentals for a large sample of US stocks in the last 10 years using a random coefficient model. Heterogeneity and omitted variable bias are properly taken into account with model coefficients being allowed to vary across time and industries. The random coefficient model allows to track waves of reliance on analysts’ forecasts and nonfundamental stock price components across time and clearly identifies the growth of the nonfundamental component in the long 1991–2000 swing.      

Mitigating political uncertainty

This study examines whether firms that establish political connections gain differential access to relevant information over legislative developments, thereby mitigating the negative consequences of uncertainty. I find that political connections (partially) offset the negative relation between investment and political uncertainty documented in prior research. My results do not appear to be driven by connected firms pursuing investments that are insensitive to uncertainty. I perform additional tests to address concerns over correlated omitted variables. First, I identify a setting around a tax policy development designed to provide new investment incentives to firms. In this setting, I predict and find that reduced information asymmetry for connected firms results in delaying investment in anticipation of future lucrative tax incentives. Second, I perform a falsification test and document that political connections do not mitigate the effects of general economic uncertainty. Finally, I continue to find support for my hypothesis within a propensity matched sample.     

Interventions in Markets with Adverse Selection: Implications for Discount Window Stigma

I study the implications for central bank discount window stigma of a workhorse model of adverse selection in financial markets. In the model, firms (banks) need to borrow to finance a productive project. There is limited liability and firms have private information about their ability to repay their debts, which gives rise to the possibility of adverse selection. The central bank can ameliorate the impact of adverse selection by lending to firms. Discount window borrowing is observable and it may be taken as a signal of firms' credit worthiness. Under some conditions, firms borrowing from the discount window may pay higher interest rates to borrow in the market, a phenomenon often associated with the presence of stigma. I discuss these and other outcomes in detail and what they suggest about the relevance of stigma as an empirical phenomenon.     

Raising capital after IPO withdrawal

We document outcomes for withdrawn U.S. IPOs (1206) from 1998 to 2017. While prior literature has studied post-withdrawal IPO re-filings and M&A, we consider private capital raising as an alternative to these outcomes as well as a potential complement. We identify factors driving private capital raising post withdrawal. We also examine the impact of key regulatory changes enacted over our sample period intended to ease capital formation. We find no significant impact of the adoption of SEC Rule 155 but a positive effect of the JOBS Act. Private capital raising post withdrawal supports second time IPO plans but not M&A.