Thread: E-Verify Isn't the Answer

E-Verify Isn't the Answer



A bill that would require employers nationwide to use the Department of Homeland Secutity’s E-Verify system to authenticate an individuals legal work status has has made it through the judiciary committee. To many, applying information technology to what seems like a nice, tidy identity problem seems like a quick, easy fix. But, this is not a simple identity problem.

A 2009 independent report (PDF) by Westat of Rockville MD shows that the system is 99.2% accurate. That sounds pretty good, if we’re talking about shooting percentages or exam grades. But there are systems where 99.2% isn’t good enough. I like my bank to be much more accurate than that, for example.

Here’s what the 99.2% accuracy rate means for E-Verify. US employers hire approximately 60 million workers each year. A 1% error rate (rounding) means that 600,000 legal workers will be mistakenly classified as not qualified to work. Once the system misclassifies you, the onus is on you to correct the problem. You are guilty until you prove that you’re innocent, so to speak. The study found that 22% of peopled who were classified spent $50 to correct the problem and 13% spent over $100. Half of misclassified workers missed work to correct the problem. These numbers might get better with time as more and more workers correct problems with their records, but I’m skeptical. There will always be documentation mistakes.

The accuracy isn’t likely to get better without significant changes to the US government identity regime. One percent is pretty good for a system based on names and social security numbers. Increasing the accuracy would require moving to something like biometric IDs for everyone. India’s doing it, but I doubt the US is ready to that step yet and even if we were, it has it’s own accuracy problems.

I can hear some people saying “But that’s a small price to pay to keep illegal immigrants from stealing jobs from Americans!!!â€

E-Verify is a great start on the problem BUT Lamar SMITH'S BILL MUST CRASH & BURN OVER THE PREEMPTION CLAUSE.

US citizens deserve to have discreprancies in their Social Security payments rectified, so that their bona fide contributions are accurately accounted for. There are apparently millions of individuals adversely affected by inaccurate identification. E Verify can help to identify those discreprancies and ensure that Social Security contributions are properly accounted.

These numbers don't seem to add up. How can E-Verify be 99.2% accurate when according to the article the MAJORITY of illegals screened by it pass through the system? On top of that many Americans would be misidentified as illegals. All of those people misidentified only add up to .08%? I don't buy it.

Its tough to run this study through hard numbers because of the small amount of businesses that use E-Verify, and the lack of hard numbers on illegal immigrants in the nation as a whole. But if E-Verify is as unreliable as the article says with regard to illegals, and would be detrimental to as many American citizens as the article says, its hard for me to believe that numbers so high would only add up to .08%.

I see the fatal flaw of this study now. The author claims that the 1% inaccuracy rate (rounding up) would mean that 600,000 Americans a year would be misidentified. Which means that he is assuming that the inaccuracy is purely driven by misidentified Americans.

THEN when attempting to defeat the oppositions counter-argument about cost/benefit of saving American jobs, he says that "the problems are on the other side of the accuracy issue." Shifting his argument to the supposedly high rate of illegals who pass through the system.

While attempting to make one point, he has ruined the other. Both points can not possibly be true.

He is either using poor reasoning, or being intentionally deceptive. He is either hugely over-inflating the number of Americans who would be negatively affected, or the 54% number of illegals passing through the system is totally inaccurate.

I think a little of both.

Isn't 1% of 1 million 10,000 ? It seems like their math is way off .

I am concerned about this 54% of illegals getting through the system. Well if that is true, then it would pick up about 15 million illegals, right?

When someone uses someone else's ssn, they must use that person's name and be the same gender and age. Then we would have two people working with the same ssn. So it sounds to me, they must contact all the people with that ssn and notify them. Something the ssa has refused to do.

I think it could be fixed to catch more illegals, if they wanted to. I think the program is a good start but the premption clause needs to go (because we know the federal government does not want to enforce the law) and I called the Judiciary and told them so.

It is kind of like us picking up illegals so ice can release them. Dont put it soley into the feds hands, they are corrupt.

These numbers don't seem to add up. How can E-Verify be 99.2% accurate when according to the article the MAJORITY of illegals screened by it pass through the system? On top of that many Americans would be misidentified as illegals. All of those people misidentified only add up to .08%? I don't buy it.

Figures lie and liars figure. One comment above stated that a 1 per cent error rate would mean 600,000 LEGAL workers would be put through unnecessary bureaucratic gyrations in order to prove their legality. What was not said is that it could miss up to 600,000 ILLEGAL immigrants. An error is an error, regardless of which side of the equation.

E-Verify may not be perfect but so far it looks like the best tool available. As another commentor noted above, it could highlight and perhaps lead to the identification and correction of problems with the Fed Social Security Data Base (how many times have we heard of payments going to people who have been dead for years, and others being declared when they are still alive and paying taxes?)

Also, shortcomings with E-Verify could also be identified and corrected, especially with a national system . With separate states currently using the system, lack of standardization could be a large contributor to the current error rate. A 1% (or in this case a .08%) error rate is not acceptable, but that is no reason to not use the system ID the causes of the errors, fix those, modify the system as experience identifies ways to implement improvements.