How does the IRS know that two or more companies are related as one employer?

[Peter Gulia]

To define an employer that might have enough full-time-equivalent employees that it might incur a 'play-or-pay' excise tax on not offering health coverage, Internal Revenue Code section 4980Hc(2)C(i) provides that "[a]ll persons treated as a single employer under subsection (b), c, (m), or (o) of section 414 ... shall be treated as [one] employer."

Imagine that there are six non-natural persons (a mix of S corporations, limited-liability companies, and limited partnerships) that are treated as one employer. None of these organizations uses a common paymaster or shares an EIN with any other. None of these organizations combines its Federal income tax return with any other. None of the five flow-through owners makes a personal income tax return with any other.

Together, the six organizations have 73 full-time employees. But none has more than 14 employees.

Corporation Alpha, which has nine employees, does not offer health coverage to anyone. (If it helps, organizations B, C, D, E, and F also don't offer health coverage to anyone.) Mary, an Alpha employee, gets a tax credit or cost-sharing reduction that subsidizes her Exchange-bought health insurance.

Imagine that the Exchange application that Mary completed asked her for her estimate of how many employees her employer has, and she answered what she knew - nine.

Unless the Internal Revenue Service has an amazing relational database, does the IRS lack a practical means to assess the excise tax in circumstances like these?

Am I just being stupid, or is there a gap in the Government's ability to enforce the 'play-or-pay' idea?

[masteff]

It would certainly take some complex data mining for the IRS to figure it out. An easy cross-check is to run Mary's W-2 against payroll data for that employer and see if the number is reasonable. The wild and crazy cross-check would be to get the employer's EIN from the payroll data and then run it for K-1's and 1099-DIV's and look to see if any two reciepents of those had other K-1's or 1099-DIV's in common.

I think the actual risk of discovery would be during an audit of the employer and the Service then doing the K-1/1099-DIV search. But I have no idea if that's part of their audit protocol.

[GBurns]

They figured it out:

http://www.treasury.gov/resource-center/tax-policy/tax-analysis/Documents/OTA-T2011-04-Small-Business-Methodology-Aug-8-2011.pdfhttp://www.treasury.gov/resource-center/tax-policy/tax-analysis/Documents/OTA-T2011-04-Small-Business-Methodology-Aug-8-2011.pdf

[Peter Gulia]

GBurns, thank you for reminding us about this Office of Tax Analysis paper.

While it describes some methods that the IRS could use to track from a business to one of its owners, and to find all businesses that tie to the same one owner, the paper doesn't really describe how the IRS would discover that two or more business organizations that don't have a majority owner in common nonetheless have sufficient common ownership among five or fewer persons that the organizations count as one employer.

One wonders how recently a government agency has done a study on the number of organizations that have fewer than 50 employees. Even harder would be guessing how many of those organizations are part of an employer group that has 50 employees.

[masteff]

One wonders how recently a government agency has done a study on the number of organizations that have fewer than 50 employees.

http://www.census.gov/econ/smallbus.html

See also: http://www.bls.gov/ and http://www.census.gov/econ/census/

But that still doesn't answer the employer group question.

[GBurns]

It seems common practice for S Corps to list their major shareholders as Directors and Officers on their state Corporation Annual Report. A similar listing of LLC and Partnerhips is also done. It should be easy to sort/merge to ind employer groups. There are some data mining cmpanies that are already offering services showing overlapping business relationshi using this same database. See http://www.corporationwiki.com for an example of these visualisation services.