The Accuracy of Multiyear Estimates vs. Annual Estimates for R&D

• Activity Number: 275
• Type: Contributed
• Date/Time: Tuesday, August 10, 2004 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economics Statistics Section
• Author(s): Jinhee Yang*+ and Wendy Rotz
• Companies: Ernst & Young LLP and Ernst & Young LLP
• Keywords: sampling ; tax ; simulation ; business

Abstract

Sampling is used to estimate a corporation's tax credit for qualifying research and development (R&D) expenditures. Studies encompass multiple year periods, but tax is reported on an annual basis and qualification for the credit is based on annual qualifying expenditures exceeding a base amount. When sampling multiple years, the estimated overall qualifying percentage is applied to each year's expenditures to obtain annual qualifying amounts. However, an IRS concern is that a year with qualifying expenditures below the base amount may be allowed an R&D credit because that year's low expenditures are masked by other highly qualifying years. The effect of a low year is unclear because a low year would also lower the overall qualifying percentage for the time frame, causing less credit in highly qualifying years, which may balance credit erroneously taken in a low year. This paper studies estimation of the R&D tax credit in multiple year studies with variable annual qualifying percentages. We explore whether an overall estimated credit for a multiyear time period can be reasonably estimated from an overall qualifying percentage, or whether separate annual estimates are necessary.