The auditor sampled one year, found a vendor's mistake, and multiplied it across six years — can we make the Department use the actual invoices instead of an extrapolated estimate?

Plain-English summary

A contractor headquartered in another state ("State A") was audited for consumer use tax covering July 2015 through June 2021. The auditor picked calendar 2018 as the sample year, found that one supplier (Vendor A) had been erroneously charging State A sales tax on materials delivered to Virginia job sites, held those invoices as exceptions, computed an error factor, and extrapolated it across the entire six-year period. The contractor appealed, arguing the extrapolation overstated the real liability, and backed it up by assembling every Vendor A invoice for the whole audit period and computing a detailed liability. It asked for a detailed audit instead of the estimate.

Holding 1 — no credit for the tax wrongly paid to State A. Virginia allows a credit for tax paid to another state (23 VAC 10-210-450), but only when the other state's tax was legitimately imposed — a taxable use in the vendor's state before the property was delivered into Virginia (P.D. 00-24, P.D. 23-17). The credit "does not apply to tax erroneously charged or incorrectly paid to another state"; the remedy is a refund claim against the out-of-state seller (23 VAC 10-210-540). Because the Virginia jobs weren't exempt and Vendor A delivered to the Virginia work sites, first use occurred in Virginia and the purchases owed Virginia tax.

Holding 2 — the sample flunked its own quality tests. The taxpayer's framing (pull the Vendor A anomaly out of the sample) faced the settled rule that a transaction comes out of a sample only if it's an isolated event outside normal operations (P.D. 99-35, 07-44, 18-63, 23-101). But the Department's review found structural problems bigger than one vendor:

• The sample year wasn't representative. 2018 purchases ran 26% larger than the second-largest annualized population and 100% larger than the smallest — a 66% average variance. In P.D. 20-111, a sample representing over half of total sales was held to skew the projection; the same concern applied here.
• Few exceptions, big error factor. A relatively low number of exception invoices produced a large error factor on the largest purchase population — raising doubt the sample reflected the population. P.D. 16-90 found a detailed audit warranted where record volume was low; here a detailed audit wouldn't take significant time or resources.
• Mismatched populations. The error factor was computed on total purchases for 2018, then extrapolated against total purchases for about half the periods — but against job-sheet amounts for the rest. Using two different population bases with one error factor computed on only one of them "provokes uncertainty as to the validity of the audit result" (cf. P.D. 98-49, which accepts alternate bases like gross sales when purchase data is incomplete — but consistency matters).

The taxpayer's spreadsheet wasn't the answer either. Detailing every invoice from one vendor doesn't complete the picture: exceptions in the sample imply there may be other untaxed purchases from other vendors outside the sample that an error factor is designed to capture.

Result: detailed audit ordered — cuts both ways. Given the wide population variances and sample deviation, the Department found it "doubtful" the extrapolation would land near a detailed audit's result, and returned the case to the audit staff for a detailed audit. The ruling warns expressly that the detailed audit may increase the liability (a further assessment would issue for periods still within the statute of limitations) or decrease it. A revised report and bill will follow, and no further interest accrues if the updated bill is paid within 30 days.

What this means for you

Contractors buying materials from out-of-state vendors

If a vendor charges its home state's sales tax on materials delivered to your Virginia job site, that tax is being charged in error — and Virginia gives no credit for it. You still owe Virginia use tax, and your only recourse for the other state's tax is a refund from the vendor. Audit exposure compounds for years before anyone notices, so check delivery terms and tax charged on multistate purchases now.

Challenging a sample audit

There are two distinct attacks, and this ruling shows both. The narrow one — "this transaction was isolated, pull it out" — requires proving the event was outside normal operations. The broad one — "this sample can't represent the population" — looks at objective markers: how the sample period's volume compares to other periods, how many exceptions drove the error factor, and whether the extrapolation base matches the base the error factor was computed on. Population variance data (here 26%/100%/66%) is exactly the kind of evidence that moves the Department.

Be careful what you ask for

A detailed audit replaces an estimate with reality — in both directions. The Department granted the taxpayer's requested method while warning the bill may go up, with a fresh assessment for any open periods. Weigh that risk before demanding a do-over, and note the 30-day payment window that stops further interest.

Common questions

Q: The vendor already charged us sales tax — why does Virginia get to tax the same purchase?
A: Because the other state's tax wasn't legitimately due: the materials were delivered to Virginia, so first use — and the taxable event — happened in Virginia. Virginia credits only tax properly owed to the origin state before delivery (23 VAC 10-210-450); erroneously charged tax must be recovered from the seller (23 VAC 10-210-540).

Q: Why wasn't detailing all of Vendor A's invoices enough to fix the audit?
A: Because the sample's purpose is to estimate errors across all vendors. Finding exceptions in the sample suggests other vendors' untaxed purchases may exist outside the sample too, so a one-vendor detail doesn't establish the true liability.

Q: What convinced the Department the sample was invalid?
A: Three things together: the 2018 sample year's purchase volume far exceeded the other years (26% over the next largest, 100% over the smallest, 66% average variance); a low count of exception invoices generated a large error factor; and the error factor was extrapolated against two different population bases (total purchases for some periods, job-sheet amounts for others) despite being computed on only one.

Q: Is the taxpayer's assessment reduced now?
A: Not yet — and maybe not at all. The case went back for a detailed audit, which can raise or lower the liability. The Department will issue a revised report and bill; paying within 30 days of the updated bill stops further interest.

Citations and references

Statutes and regulations:

• Va. Code § 58.1-1821 — administrative appeal (application for correction) to the Tax Commissioner
• 23 VAC 10-210-450 — credit for sales or use tax paid to another state (legitimately imposed, pre-delivery)
• 23 VAC 10-210-540 — no credit for erroneously charged out-of-state tax; refund lies against the seller

Authorities the Department relied on (described here, not linked): P.D. 00-24 (3/8/2000) and P.D. 23-17 (2/21/2023) (credit limited to tax legitimately imposed by the origin state before delivery to Virginia); P.D. 99-35 (3/29/1999), P.D. 07-44 (4/26/2007), P.D. 18-63 (5/2/2018), and P.D. 23-101 (8/24/2023) (removal from a sample requires an isolated, non-recurring event); P.D. 16-90 (5/19/2016) (low record volume can warrant a detailed audit); P.D. 20-111 (6/30/2020) (an oversized sample skews the error-rate projection); P.D. 98-49 (3/11/1998) (alternate extrapolation bases where purchase data is incomplete).

Source

• Landing page: Virginia Laws, Rules & Decisions
• Ruling: P.D. 25-35

Original ruling text

March 14, 2025

Re: § 58.1-1821 Appeal: Consumer Use Tax

Dear *:

This will respond to your letter in which you seek correction of the retail sales and use tax assessment issued to * (the “Taxpayer”) for the period July 2015 through June 2021.

FACTS

An audit was conducted on the books and records of the Taxpayer, a contractor headquartered in * (State A), for the period at issue. The auditor selected the 2018 calendar year to use as a sample for the entire of audit period. Among the invoices listed as exceptions, the auditor found that one vendor (Vendor A) charged State A sales tax in error for materials delivered to and used on a Virginia job site. These invoices were held as exceptions and were used to calculate an error factor that was extrapolated over the audit period to determine the liability.

The Taxpayer filed an application for correction contending that the extrapolation of the sample error rate overstates the actual liability. To support its position, the Taxpayer gathered all of the invoices from Vendor A for the entire audit period and used them to calculate a detailed liability for the audit period. The Taxpayer requests that the Department perform a detailed audit.

ANALYSIS

Tax Paid To Another State

With regard to taxes paid to another state, Title 23 of the Virginia Administrative Code (VAC) 10-210-450 allows a credit for taxes paid elsewhere. However, such a credit is intended only to apply to taxes owed in the state from which the property was purchased, if legitimately imposed because of a taxable use made in the vendor’s state, and prior to the delivery of the property in Virginia. See P.D. 00-24 (3/8/2000) and P.D. 23-17 (2/21/2023). Title 23 VAC 10-210-540 states, in part, that:

This credit does not apply to tax erroneously charged or incorrectly paid to another state. For example, if a person purchases and takes delivery in Virginia of tangible personal property purchased from an out-of-state dealer who incorrectly charges out-of-state tax, no credit is available. The purchaser must apply to the out-of-state seller for refund.

In this case, the Taxpayer’s jobs in Virginia were not tax exempt. Further, it appears that Vendor A was delivering purchases to the Taxpayer’s work sites in Virginia. Under such circumstances, the first use of the property by the Taxpayer occurred in Virginia and the transactions were subject to Virginia sales and use tax.

Sampling

The Department’s audit utilized sampling to determine the amount of assessed liability. Sampling is an audit technique of significant value that is widely used in both the public and private sectors for all types of audits where a detailed audit would not prove beneficial either to the auditor or the client. When sampling techniques are properly applied, the final results are usually within a narrow percentage range of the actual amount that would have been determined by a detailed audit.

The purpose of the audit sample is to determine a factor for errors within a representative select period. Once the error factor is determined, the factor is extrapolated over the entire audit period. The purpose of the projection is to account for likely similar transactions on which Virginia tax has not been paid. Every effort is made to select objectively the sample periods that are representative of the period being audited.

The auditor explained the sampling procedure and the Taxpayer agreed to utilize sampling at the outset of the audit. The Taxpayer contends that the sampling technique utilized is improper in this instance because one vendor incorrectly charged tax to State A during the sample period distorting the error factor. The Taxpayer requests that the Department use the additional invoices and spreadsheet provided to detail the precise amount of tax due rather than using the estimate established by sample’s error factor.

In order for a transaction to be removed from the audit sample and the extrapolation, the Taxpayer must establish that the transaction is an isolated event and not a part of its normal operations. See Public Document (P.D.) 99-35 (3/29/1999), P.D. 07-44 (4/26/2007), P.D. 18-63 (5/2/2018) and P.D. 23-101 (8/24/2023).

While block sampling can be a prudent way to determine a liability when the volume of transactions is great, a relatively low number of invoices were listed as exceptions in this audit of the Taxpayer. Further, the sample resulted in a large error factor for the largest purchase population. These factors raise concerns as to whether the sample is representative of the population.

In P.D. 16-90 (5/19/2016), the Department determined that a low volume of records warranted a detailed audit of a dealer. Because of the low volume of records that were reviewed for the sample period (the largest population period within the audit period) and the potential that even fewer documents would need to be reviewed for other population periods, the performance of a detailed audit under these circumstances does not appear to entail a significant amount of audit time or resources.

As indicated above, the purchases for the 2018 tax year were substantially higher than any other year in the audit period. The 2018 population was 26% larger than the second largest annualized population and 100% larger than the smallest population. The average population variance for between the 2018 population and the remaining annualized populations was 66%.

In P.D. 20-111 (6/30/2020), the Department determined that the inclusion of a sample that represented over 50% of the total sales for the remaining sample months skewed the error rate projection over the population. After reviewing the sample and exceptions list in this case, it appears that the purchases for the 2018 tax year may not be representative of the Taxpayer’s normal business activity during the audit period.

In addition, according to the audit report, the error factor was computed based on total purchases for 2018. The error factor was extrapolated against total purchases for about half of the periods under audit. However, the remaining periods were computed by applying the error factor for total purchases to amounts reported on job sheets. The use of two different populations in a sample base with a single error factor calculated on only one of the population bases provokes uncertainty as to the validity of the audit result.

In P.D. 98-49 (3/11/1998), the Department addressed the issue of incomplete purchase data. In conducting an audit of purchases, a dealer or consumer may find it difficult to provide purchase information on which to extrapolate the results of a sample. In such circumstance, gross sales would be an acceptable audit procedure to extrapolate the results of a purchase sample. In this case, it does not appear consideration was given to the differing sample base populations.

The additional information provided by the Taxpayer also fails to provide a complete analysis of the potential audit liability. The fact that exceptions were found in the sample period means there may be other untaxed purchases from other vendors outside of the sample period that are not included in the error factor. Thus providing the detail for all transactions for one vendor during the audit period is not sufficient to alter an audit finding.

DETERMINATION

Given the wide variances in the populations and deviation of the sample, the chances that the final results fall within a narrow percentage range of the actual amount compared to a detailed audit appear doubtful. Accordingly, the case will be returned to the audit staff to conduct a detailed audit.

In changing from a sample audit to a detailed audit of sales, it is possible that the result may increase the Taxpayer's liability. In the event that an increase in liability is found, the Department would revise the audit and issue another assessment for the additional liability found for the audit periods still within the statute of limitations for assessing the tax. The audit assessments will be adjusted appropriately if a decrease in liability occurs.

After the revision of the audit is complete, we will issue a revised audit report and revised bill, if applicable, with interest accrued to date, to the taxpayer. No further interest will accrue provided the outstanding liability is paid within 30 days of the date of the updated bill.

The Code of Virginia section and regulation cited are available online at law.lis.virginia.gov. The public documents cited are available at tax.virginia.gov in the Laws, Rules, & Decisions section of the Department’s website. If you have any questions regarding this determination, you may contact * in the Office of Tax Policy, Appeals and Rulings, at or **.

Sincerely,

James J. Alex

Tax Commissioner

Commonwealth of Virginia

AR/4401.B

Related Documents

98-49

99-35

00-24

07-44

16-90

18-63

20-111

23-17

23-101