Correcting for Cross-Sectional and Time-Series Dependence in Accounting Research

Abstract

We review and evaluate the methods commonly used in the accounting literature to correct for cross-sectional and time-series dependence. While much of the accounting literature studies settings in which variables are cross-sectionally and serially correlated, we find that the extant methods are not robust to both forms of dependence. Contrary to claims in the literature, we find that the Z2 statistic and Newey-West corrected Fama-MacBeth standard errors do not correct for both cross-sectional and time-series dependence. We show that extant methods produce misspecified test statistics in common accounting research settings, and that correcting for both forms of dependence substantially alters inferences reported in the literature. Specifically, several findings in the implied cost of equity capital literature, the cost of debt literature, and the conservatism literature appear not to be robust to the use of well-specified test statistics.

A comparison of the value relevance of interim and annual financial statements

“…A 3-month lag period was used to allow time for accounting information to be communicated to the public. Following Chen, Liu and Ryan (2008), Gow, Ormazabal and Taylor (2010) and Peterson (2009), we corrected for cross-sectional and time-series dependence in our data by clustering standard errors by firm and by year. Peterson (2009) argues that clustering standard errors on multiple dimensions (by firm and by year) will yield correct inferences and correct for correlated residuals in the dataset.…”

“…***, ** and * denote significance at the 1%, 5% and 10% level, respectively. (Gow et al 2010) of Equation 1. The sample consists of a pooled dataset (annual and interim data combined into one dataset) for annual and interim accounting data.…”

“…Table 4 contains the results of the OLS pooled regression, testing whether accounting information published in annual and interim financial statements is value relevant (step 1 to test hypothesis 3). As our data are subject to cross-sectional and time-series dependence, we report results for the coefficient estimates and t-values that are calculated using firm and time-clustered standard errors (Chen et al 2008;Gow et al 2010;Peterson 2009). …”

The Effect of Matching on Firm Earnings Components

“…To investigate the relation between the persistence of firm earnings components and matching, we utilize the following OLS regressions that model one-year ahead earnings (cash flows) as a function of industry-wide and firm-specific earnings components. Following Gow et al (2010), we use the two-way clustered standard errors (by firm and year) to report the p-values, controlling for heteroscedasticity and correlation among our firm-year observations.…”

The Impact of OECD Countries’ Macroeconomic Factors on Turkey’s Foreign Trade

“…Petersen (2009) compared these resistive standard errors, and proposed to use two-way clustered resistive standard errors for resistivity control. Gow, Ormazabal and Taylor (2010) found that two-way clustered resistive standard errors are required for a valid deduction in many accounting practices. On the other hand, two-way clustering method can be used for only specific and limited error structures .…”