Page 61 - TaxAdviser_2022

TAX CLINIC

business components under develop- the four-part test is required to be at qualifying percentages. These prelimi-
ment during the tax year is reviewed to the business-component level. When nary qualifying percentages can often
identify only those business components developing a statistical sampling meth- be leveraged to create smaller sample
with qualified activities and QREs. The odology, taxpayers should first attempt sizes. However, taxpayers are still re-
revised population is next divided into to use business components as the quired to analyze the four-part test
several strata (e.g., buckets) of similar sampling unit because this most directly and substantiate such analysis at the
types and sizes, and random samples are aligns with the requirements of Sec. business-component level under this
selected from each stratum. Tax research 41(d)(2)(A) and Regs. Sec. 1.41-4(b)(1) sampling approach.
credit professionals must subsequently and the substantiation requirements of Again, taxpayers must use caution to
analyze the samples from each stratum Regs. Sec. 1.41-4(d). Often, projects implement this kind of study properly.
in accordance with Sec. 41 and the directly align with business components Taxpayers must demonstrate nexus
regulations thereunder to determine the and are easily listed, and projects are between employee time and qualified
QREs of each sample. After QREs have therefore a natural sampling unit chosen business components for each randomly
been determined and substantiated for by many taxpayers. sampled cost center. A two-stage sam-
each sample, statisticians use the analysis For various reasons, including ease of pling approach can be applied to reduce
results to estimate total QREs for the data collection, taxpayers often establish the burden of demonstrating nexus for
entire population list. sampling units as something other than each cost center. Under this approach,
During the sampling process, statisti- business components. For example, a random sample of business compo-
cians will calculate the accuracy of the sampling by employee may be a more nents is analyzed for each randomly
estimates, as measured by confidence practical choice for taxpayers that do not sampled cost center or department.
and precision. Broadly, confidence repre- capture employee time via time-tracking To maintain statistical validity, the as-
sents the chance that the selected sample software. However, taxpayers must use sessed business components should
will produce an estimate close to the caution when using non–business- not be determined based on qualifying
true value (i.e., total QREs of the entire component sampling units. To properly activities and should instead be gener-
population), and precision indicates the substantiate the research credit under ated based on valid, statistical random
proximity of the estimate to the true this sampling approach, taxpayers must samples. In general, other non–business-
value. The IRS specifies confidence and generally demonstrate nexus between component sampling (e.g., location,
precision requirements in Rev. Proc. QREs for each randomly sampled supervisor, or a combination) follows the
2011-42, and taxpayers should col- employee and qualified business com- same methodology.
laborate with an experienced statistician ponents that meet the four-part test
knowledgeable of IRS statistical expec- requirement. Taxpayers must track and Choosing the right sampling
tations when sampling. document the complete list of business method
components on which each randomly The above-described statistical sam-
Constructing the sample and sampled employee worked to appro- pling methodologies are some common
substantiation requirements priately establish nexus. Taxpayers may approaches to statistical sampling for
The first step in the statistical sampling choose to use a two-stage sampling research credit studies. However, there
process is to construct the population methodology whereby business compo- are many other acceptable methods for
listing or sample frame (i.e., listing of nents are randomly selected for the ran- statistical sampling, including variations
every item in the scope of the study). domly selected employees to minimize in sampling units. Taxpayer-specific facts
Building the sample frame requires demanding documentation require- including types of research performed,
expenses to be aligned with a sampling ments. Alternatively, taxpayers with few data availability, and organization struc-
unit (e.g., a row of data in the popula- business components and/or few em- ture should be considered to develop
tion sampling frame data set listing). ployees may find it practical to survey all the most appropriate statistical sam-
Common sampling units in research employees who do not use time tracking pling methods.
credit studies include business compo- at the business-component level. Statistical sampling is an IRS-
nents, projects, employees, cost centers, Sampling by cost center or depart- accepted approach that offers several
locations, departments, supervisors, or a ment can be highly efficient if per- analysis- and documentation-related
combination of these. formed properly and adequate data is benefits that taxpayers can implement to
Guidance provided in the FAA available. For example, statisticians may compute the research credit. Further, sta-
amplifies the critical nature of determin- use historical cost center or department tistical sampling may prove to be a more
ing sampling units, as application of information to generate preliminary efficient methodology for computing

14 February 2022 The Tax Adviser