Michael Ettredge (University of Kansas, School of Business), Lili Sun (Rutgers University (Newark), Business School),  Picheng Lee (Pace University, Luban Business School) & Asokan Anandarajan (New Jersey Institute of Technology, School of Management) have posted Using Deferred Tax Data to Detect Fraud on SSRN.  Here is the abstract:

The objective of this paper is to examine whether deferred tax data potentially can be used to develop red flag signals of earnings-overstatement fraud. We use a sample of 105 firms sanctioned by the SEC (in Accounting and Auditing Enforcement Releases, AAERs) for earnings overstatement fraud, and a sample of 105 control firms matched by year, asset size, and two-digit SIC code. Tax variables examined include book income minus taxable income (BMT), the change in BMT from year-to-year (BMTCHG), deferred tax expense (DTE), and the change in DTE from year-to-year (DTECHG). Each variable is scaled. Our results indicate that among firms reporting positive pretax book income, BMT and DTE are positively and significantly associated with the occurrence of fraud. When both BMT and DTE appear simultaneously in the model, both variables retain significance, and BMT is more significant. Although tax-change-related variables BMTCHG and DTECHG are insignificant, models incorporating their interactions with an indicator for positive book income have significantly better overall fit than the base model. This suggests that BMTCHG and DTECHG also potentially provide evidence of the occurrence of fraud, although they do not perform as well as BMT and DTE. In summary, this study identifies several new tax-related variables that potentially can be used as red flags for fraud detection, and which are not currently in the red flag checklists recommended to auditors by Statement of Auditing Standards No. 99. The findings of this study are also potentially applicable in Canada and internationally since Canadian accounting standards related to deferred taxes are similar to SFAS No. 109, and neither Canadian nor international auditing standards suggest tax-related variables as red flags for fraud.