Now that the government shutdown is over, Congress’ attention will turn to other issues. There is a possibility that a series of immigration reform bills will be voted on in the House of Representatives before the end of the year. One bill will certainly include mandatory E‑Verify.

As my colleagues and I have written over the last several years, E‑Verify is bad for businesses, taxpayers, the privacy of all Americans and residents, economic growth in general, and it won’t stop unlawful hiring. Don’t believe me on the last point? Just look at Arizona. Here is a table of the number of all new hires in the state and the number of E‑Verify queries run per quarter:

Source: U.S. Census and Department of Homeland Security

Although all hires in Arizona are supposed to be run through E‑Verify, an average of just over 50 percent of hires actually were from 2008 to the end of 2012. These numbers actually overstate E‑Verify’s enforcement record because multiple E‑Verify queries could be run on the same hire. The numerator could be a lot smaller than is reported above.

If a state like Arizona will not enforce E‑Verify, what chance is there that the federal government will do it everywhere? Thankfully, lax enforcement of E‑Verify in Arizona is a good indicator that this harmful system will not get the chance to be as destructive as many of us fear if it is ever mandated nationally.

Despite these terrible numbers, some continue to push E‑Verify as a magic bullet that will put an end to the hiring of unauthorized immigrants and make our country’s byzantine and restrictive immigration laws enforceable.

A major problem with E‑Verify is that the system is unreliable. The Westat report, the last major study of E‑Verify, noted that the tentative non‐​confirmation rate (TNC) error rate declined from 0.7 percent to 0.3 percent from 2009 to 2012. TNCs are issued when a prospective employee’s information raises a red‐​flag with the system. On the surface, the low rate seems impressive and the improvement is certainly commendable.

But with a pool of 10,000 employees, that error rate would produce 30 erroneous TNCs. Apply those rates to a mandatory, national pool of 150,000,000 workers and it would result in 450,000 erroneous TNCs. That is a substantial number of legal workers who would be forced to jump through regulatory hoops while exposing themselves and their employers to additional costs in order to prove their legal status and get permission from the government to work.

Similarly, the Westat report notes that final non‐​confirmations (FNCs are issued when a TNC appeal fails) are issued correctly 94 percent of the time, with 6 percent issued wrongly to those who would otherwise be eligible to work. The report noted that, for FY2009 (the year of data used by the report), 190,165 total FNCs were issued, with 178,193 (approximately 94 percent) issued correctly, and 11,972 (approximately 6 percent) issued incorrectly to legal workers.

Applied to the national pool at large, the numbers would be huge. One estimate puts potential job losses from FNC problems in a national system at a maximum of 777,000 under worst case scenario. Those numbers should give E‑Verify’s cheerleaders some pause.

When E‑Verify shut down along with the government, employers enrolled in the system relied on the I‑9 as an alternative way to verify employee work eligibility. The Legal Workforce Act would mandate E‑Verify if it passes but it would also eliminate the I‑9 requirement – likely the only good provision of that bill. But if this bill becomes law and the government shuts down at any time in the future, there would be no way for them to verify the work eligibility of employees until the government reopens and a backlogged agency issues new rules.

E‑Verify will not stop the hiring of unauthorized immigrants and will not be an enforcement magic bullet despite the technocratic wishes of its supporters. What it will do, however, is put legal American workers through the proverbial wringer and likely force enormous numbers of currently legal hires into the black market – just as it has in Arizona. That is not a preferable policy outcome.