Risk Factor iconRisk Factor

Obamacare Exchange Sign-ups Hobbled by IT Systems Not Ready for Prime Time

I don’t need to tell anyone about the controversy surrounding the Affordable Care Act (more commonly known as Obamacare). It was the central issue in the game of brinksmanship that led to the U.S. government shutdown last week. But mirroring that mind-blowing dysfunction was the less-than-stellar 1 October rollout of the federal website healthcare.gov. The Obamacare-mandated Web portal lets consumers who don’t have employer-sponsored medical insurance meet the legal requirement to sign up for health coverage through the states where they reside. (To be precise, healthcare.gov is for residents of 36 states whose governments opted not to set up independent healthcare exchanges.) Online exchanges for some of the other 14 states and the District of Columbia also debuted with disappointing results.

So, what happened? Well, it’s no secret that governments are terrible at IT project implementation. Examples abound—as regular readers of The Risk Factor are well aware. (Some of our reporting on recent foul-ups is here, here, and here.) There’s been little evidence so far that these projects are any different.

The Department of Health and Human Services (HHS) reported Wednesday that there were 6.1 million unique visitors to healthcare.gov on the first day and a half after the site opened on Tuesday. By Friday, that number had surpassed the 8 million mark. That’s a good indicator of the level of interest in getting signed up for health coverage. But it’s only part of the picture. What HHS purposely left out (and left to our imagination) is the actual number of enrollments. Officials said they would probably release enrollment numbers next month after tabulating totals from, online, call centers, and paper enrollments. But the picture that’s forming based on anecdotal evidence is not pretty.

Most attempts to reach the federal website resulted in this:

“We have a lot of visitors on our site right now, and we're working to make your experience here better. Please wait here until we send you to the log-in page. Thank you for your patience.”

Or this:

“Important: Your account couldnt (sic) be created at this time. The system is unavailable."

According to a Los Angeles Times story, community groups aiming to help people sign up have been frustrated in their attempts to do so. Even large insurance companies, which have a vested interest in getting people enrolled in the exchanges, were unsuccessful in the early going. For example, a spokesman for Blue Cross Blue Shield of Louisiana, that state's largest insurer, told the Los Angeles Times that, as of Wednesday, the company hadn't been able to enroll anyone through the federal website. Others who left in frustration included reporters including one for the Huffington Post, who said: “Though officials from the Centers for Medicare and Medicaid Services said they'd made strides correcting the federal exchanges' problems, The Huffington Post made dozens of attempts and still couldn't sign into the website late Tuesday afternoon.”

"We have had a few slowdowns, a few glitches, but it's sort of a great problem to have. It's based on the fact that the volume has been so high and the interest is so high," Health and Human Services Secretary Kathleen Sebelius said on MSNBC Tuesday. "We're working quickly to fix that."

U.S. Chief Technology Officer Todd Park explained further, pointing out that the government expected HealthCare.gov to draw only as many as 60 000 simultaneous users. That estimate was apparently based on a projection from the volume experienced nearly a decade ago on a site for Medicare Part D. But at peak, the Obamacare site was being deluged by up to 250 000 people at a time.

"These bugs were functions of volume,'' Park told USA Today. "Take away the volume and it works.''  Right. Take away widespread interest in signing up for health insurance, and the portal through which people are supposed to sign up for health insurance will work as intended.

The system’s performance invited a swipe from an IT official from the previous administration. “Whoever thought it would draw 60,000 people wasn't reading the administration's press releases,” David Brailer, former national coordinator of health care information technology under George W. Bush, told USA Today. “The Medicare Part D site [launched in 2006] was supposed to have 20,000 simultaneous users and was [built to accommodate] 150,000, and that was back when computing was done on an abacus. It isn't that hard.”

The news wasn’t any better with the state-run exchanges. California residents were stuck in traffic along both routes to enrollment there: computer glitches stymied attempts to sign up online, while hold times at telephone call centers topped 30 minutes. The computer system created to, among other things, log a consumer’s data and determine whether he or she is eligible for government subsidies to cover part of the premiums, responded so poorly that its operators were forced to shut down the online enrollment system twice. According to the L.A. Times story, “Officials were pleased with the strong consumer interest and vowed to fix the problems.”

On the opposite coast, officials in the second most populous state fielding its own exchange reported what could generously be described as an anomaly. State of Health, the healthcare portal serving New York State, which has a population of roughly 18 million, had reportedly received 30 million hits by late Wednesday, prompting some observers to suspect that hackers may plotting a break-in or an out and out takeover. Whether that’s true or not, Donna Frescatore, director of the state’s exchange, confirmed that despite all that activity, only about 12 000 people had managed to enroll by Wednesday evening.

Responding to questions about the extraordinarily high volume, Frescatore told the Wall Street Journal that, “We have no evidence that this is anything but people learning more about [the site].” Furthermore, said Frescatore, state officials are not looking into the possibility that cybercrime was a contributing factor.

Ahhh…the power of positive thinking.

We can all keep our fingers crossed, but the issue of security will likely pop up again. As we recently reported on this blog, privacy safeguards have likely taken a backseat to getting the exchanges open on time. Another IEEE Spectrum post focusing on the exchanges’ security issues is here.

There’s no question that the overwhelming interest caught New York flat footed. Officials took the Web portal offline Tuesday night. Once the smoke cleared, they doubled its capacity and implemented some fixes aimed at keeping it from getting hung up as it did throughout its first day of real-world operation. What happened on day two? The same thing, more or less.

California and New York weren’t alone in their misery. According to a Huffington Post article, at “Maryland Health Connection, Kynect [Kentucky’s exchange], Connect for Health Colorado, Rhode Island's Health Source RI and others, consumers faced obstacles to setting up accounts or comparing plans—or even viewing the websites at various points in the day.” The Chicago Tribune reported that a glitch affecting Illinois’ exchange—missing fields in an online form—left people attempting to enroll in on the first day unable to figure out whether they were eligible for the federal subsidy for premiums. Though that problem was remedied by the middle of the day, sailing still wasn’t smooth, said the Chicago Tribune article. Illinois Gov. Pat Quinn’s advice for those who had trouble accessing the site? "Just keep trying."

Criticism of the portals’ bumpy first week has come from all quarters. But the rollout still has its apologists. In an e-mail sent to the Huffington Post, Jonathan Gruber, a Massachusetts Institute of Technology economist who was an architect of the 2006 Massachusetts health care regime after which Obamacare was modeled, says,

"Hours or even days is not the relevant timeframe for evaluating exchanges. The question is simply whether there are ways that folks can sign up to get insurance by Jan. 1. That is a question for late November, not early October. If things are really buggy in six weeks, that could be more of an issue."

The best “Keep hope alive” message had to be the widely reported one delivered by HHS Secretary Sebelius.

“I clearly have an iPad and I also have an iPhone and about 10 days ago I got the prompt that the operating system had changed,” Sebelius said. Noting that the experience wasn’t great, she added that, “everyone just assumes ‘well there’s a problem, [Apple will] fix it.’” Here’s the good part: “We’re building a complicated piece of technology, and hopefully you’ll give us the same slack you give Apple.”

Matthew Yglesias, writing for Slate, deftly picks the Sebelius comment apart:

“Apple, like any private business, is customer-driven. Apple knows that if it doesn’t provide good products and services, the public will exercise its options, and go to Samsung and Android, or Windows, or even Blackberry…Apple, the world’s most-valuable brand, has a reputation for producing quality products that work. The government has exactly the opposite track record. There is no public confidence in government programs, whether they be in veterans’ affairs, the postal service, the stability of Social Security, containing spending, managing contracts, rooting out fraud, the IRS, the NSA, the EPA, immigration, self-investigating, protecting our Embassies and personnel — you name it.”

As the federal and state governments have repeatedly reminded us, the more than 40 million U.S. residents without employer-sponsored health insurance have until 15 December to enroll in order to get coverage on 1 January, and until 31 March to avoid being assessed a penalty. Will the sites’ managers get their respective acts together in time? I won’t call Sebelius’ and Gruber’s optimistic takes on the situation into question. I’ll simply direct readers’ attention once again to the Risk Factor links in the second paragraph of this post. They’re concrete illustrations of the points Yglesias makes.

Here are links to several other related articles:

Photo: Mike Segar/Reuters

IT Hiccups of the Week: California’s Unemployment System Upgrade Saga Continues

Last week was a good week for IT project spotters to add to their snafu collection. We review our collection’s additions this week with a system snafu in California that just keeps chugging along with no real end in sight.

California’s Unemployment Insurance Upgrade Problems Rumble On

California’s Employment Development Department (EDD) put out a press release last Friday that tried to put another heavy layer of lipstick on its modernized unemployment insurance system’s pig. The EDD stated that it is “wrapping up work on new upgrades to our 30-year-old payment processing system” and that one month after the upgrade went live, it has reached the milestone of processing “about 83% of all certifications from all claimants within the week in which the certification was received.” Of course, 17 percent out of 800 000 who are on unemployment in California is still a pretty big number.

Also, notice the use of the word "about" used in the press release. As I noted earlier this month, EDD has had repeated trouble figuring out how many unemployment claims it has failed to process correctly. The number has grown steadily; first it was only 5000, then it grew to 20 000, then 50 000, and finally, at least 185 000. What also is unclear from the EDD press release is what it means by “claims processing.” Because of problems with the system upgrade, EDD personnel have been working lots of overtime to process unemployment claims by hand. It would helpful if EDD would break out its “about 83%” figure into those performed correctly by the new upgraded system, and those it still has to process by hand.

At one point early last week, some 124 000 Californians had yet to receive their unemployment checks—many many of whom who had been waiting since early September. Now that number has been whittled down to about 40 000. However, a big reason for that drop was Gov. Jerry Brown ordering the EDD last week “to start issuing the delayed checks immediately and verify the recipient’s eligibility afterwards.” EDD says it “hopes” to send out more checks later today.

Somehow I doubt the cost (or time) of verifying eligibility after the fact was budgeted into the upgraded system’s original project budget, although it probably should have been. Why? Glad you asked.

First, state unemployment insurance upgrade projects don’t have a good development track record: just ask Colorado or Pennsylvania, both of which canceled their projects after not being able to get them to work even after massive cost overruns.

Second, even when the systems go live, they aren’t very reliable, as Nevada and Massachusetts unemployed residents have found out over the past few weeks. Unemployment insurance officials always admit that they expected problems, but they are inevitably surprised that there are many, many more of them than they expected. For instance, an EDD spokeswoman told the LA Times that the department expected some “hiccups” with the system upgrade, “but we didn't know it'd be to this magnitude.”

Third, even when a system’s performance is admittedly “unacceptable” as California’s Secretary of Labor and Workforce Development Marty Morgenstern stated last week, it doesn’t prevent other government officials in the same department from claiming that problems are really not that bad. They only appear bad, they say. In Massachusetts, for example, Labor Department Secretary Joanne Goldstein insisted last week that the state’s new $46 million unemployment insurance system was operating with minimal problems even as a flood of complaints about the system came out after its rollout three months ago, the Boston Globe reported. The state had to negotiate a contract warranty extension with Deloitte Consulting, the prime contractor, a move made necessary in order to fix those supposedly minimal problems.

Finally, government IT program managers like to live dangerously. For instance, EDD gave the contract for its unemployment system to Deloitte “despite a string of problem projects” across California state and local governments that littered the company's track record, the LA Times reported late last year. These included: a project canceled after four years by California’s Department of Developmental Services after it figured out the system it paid $5.7 million to Deloitte to develop didn’t work; a project to link California's court system computers, originally slated to cost $33 million, but canceled because it couldn’t be made to work even after Deloitte pocketed $330 million; a botched payroll system Deloitte developed for the L.A. Unified School System; and a Deloitte-led, botched ERP system development for Marin County, California, among others. California government IT has been very, very good to Deloitte's bottom line.

Massachusetts admitted a few weeks ago that its original contract with Deloitte to modernize the state's unemployment system was “flawed” and allowed Deloitte  “to miss deadlines and still charge the state some $6 million more than originally planned,” the Boston Globe reported earlier in the month. Suzanne Bump, the Labor Department Secretary in charge of the contract at the time—who is currently State Auditor—told the Globe she has “no recollection of what was in that contract language.”

It is always amazing how selective amnesia seems to strike government officials whenever they are asked to explain their unexplainable decisions.

Jeep’s Transmission Problem Fixed

Usually, U.S. auto manufacturers work overtime to ship new model year vehicles to their dealers in August in order to take advantage of the big September Labor Day sales weekend. This typically means a ramp up in production during the summer months at the automobile production plants. So it was a surprise when early last week, Chrysler announced that it was temporarily laying off the several hundred workers it added just in mid-August at its Toledo Assembly Complex where its new and highly anticipated 2014 Jeep Cherokee was being produced. Chrysler told Automotive News that it had produced a “critical number” of vehicles, and it didn’t want to place too much of a strain on its logistic partners by shipping them out.

Chrysler’s explanation didn’t make any sense to many observers because Chrysler had been insisting that it would be selling the Cherokees in volume by now. Automotive News did a check and found that “no Cherokees were listed in dealer stocks anywhere in the United States.” The only thing that Chrysler did say to add clarity to its nonsensical statement was that the Cherokees required a “software fix” before the vehicles would be sent to its dealers.

Well, on Thursday, the head scratching ended when Chrysler reversed course and announced that it would soon begin delivering the new Jeep Cherokees to dealers after all. The Wall Street Journal reported that Chrysler had been having “problems tuning [the new vehicle's] nine-speed transmission” which caused the car not to shift as smoothly as expected at different speeds and temperatures.

The Journal reported that earlier this summer, “Chrysler delayed test drives for the media on the Cherokee because it was having trouble working out all the bugs in mating the new nine-speed transmission to the vehicle's engines.” According to Chrysler, the 2014 Cherokee has “the world's first application of a highly technical nine-speed transmission” and with it being “mated to two new engines and three complex 4x4 systems,” it wasn’t surprising that technical issues arose.

The company is updating the Cherokee’s powertrain software, and hopes that dealers will be able to sell the new Cherokees very soon. Those with a long memory may remember that there was a flaw in the transmission software for Chrysler’s new 2008 Jeep Grand Cherokees and Commander SUVs that caused engine stalling. The vehicles were subject to an embarrassing recall in May 2008, something that Chrysler wants to avoid this time around.

Obamacare Public Health Exchanges Cross Their Fingers

Well, tomorrow is when Americans can start to sign up for the public health exchanges being created under the U.S. Patient Protection and Affordable Care Act (pdf). As I have noted previously, a number of states are warning that problems may await those signing up. The list got a little longer last week, when the District of Columbia announced that computer issues have caused its exchange’s opening to be delayed for an unspecified amount of time.

A story today at the New York Times provides a good overview of the situation at several state exchanges, which, in a word, can be described as “tense.” Another word that could be applied is “uncertain.” For instance, Rocky King, the executive director of Oregon’s new health insurance exchange, which has delayed the full operations of its exchange to engage in more testing, told the Times that even when the exchange goes live with limited functionality tomorrow, “We could crash and burn and have to close it down.”

No one really will know until tomorrow how many of the exchanges are ready for prime time. We’ll report on the good, bad and ugly in next week’s IT Hiccups review.

Of Other Interest ...

New York City Targets Risky Taxi Drivers after Computer Glitch

Heathrow Passengers Stranded by Baggage Glitch

Apple Maps Glitch Sends Drivers onto Alaskan Airport Runway

Google Gmail Glitch Hits Millions

Wendy's Restaurant Glitch Charges Customers Multiple Times

UK Nationwide Bank Shows Customer Accounts as Bring Empty

Orbital’s Cygnus Docks With ISS After Software Fixed

Photo: Photo: Rich Pedroncelli/AP Photo

This Week in Cybercrime: NSA Wants More Info from Firms

You have to give it to Gen. Keith Alexander, head of the U.S. National Security Agency (NSA). The man can stand up to abuse. He’s faced the ire of attendees at public events ever since the spy agency's monitoring U.S. citizens’ electronic communications was leaked earlier this year. The aftermath of Wednesday’s keynote address at the Billington Cybersecurity Summit, where he called upon the private sector to partner with the NSA, the FBI, the Department of Homeland Security, and the CIA to prevent or limit cybercrime was no different. He couldn’t possibly have expected any different after he said, "We need the authority for us to share [cyberattack information] with [private businesses] and them to share with us."

Despite revelations about the NSA’s activities—some that directly contradict previous government assurances about the limits of the surveillance programs—Alexander insisted that the NSA hasn’t done anything illegal. Furthermore, he said, the calls from some members of Congress to limit the reach of the NSA and the nation’s other spy and law enforcement agencies are based on what he calls sensationalized reporting. Alexander pushed for even more data access from U.S. companies. The more information companies shared with NSA the more cyberattack warnings it could supply to them.

But many observers now see that rationale as threadbare and view Alexander and his ilk with a jaundiced eye. Jerry Brito, a researcher who heads the Technology Policy Program at the Mercatus Center at George Mason University, in Virginia, told CSO that the NSA already has the authority to share data with companies. It could simply declassify information, allowing companies to use it to protect themselves. But that’s not what the agency is interested in, Brito insists. "What they really want is more information about the communications of Americans under the rubric of cybersecurity information sharing," he told CSO.

Stolen Data Clearinghouse Gets Info from Its Above-Board Counterparts

Willie Sutton’s famous response to being asked why he robbed banks—“Because that’s where the money is”—could certainly be the rationale behind a recently discovered cybercrime program targeting data brokerage firms. According to an investigative report [pdf] from security reporter Brian Krebs, an online identity theft service that specializes in selling Social Security numbers, credit and background check reports, and other information, gained access to the data by hacking into the networks of companies such as LexisNexis, Dun & Bradstreet, and an employment background screening company called Kroll Background America Inc. Botnets in the companies’ systems continually siphoned off information and passed it to servers controlled by the cybercrooks.

The criminal clearinghouse, whose website was at SSNDOB[dot]MS, had served some 1300 customers who paid hundreds of thousands of dollars to get their hands on the SSNs, birth dates, drivers license records, and the credit and background check information of more than four million U.S. residents.

Researchers Identify Source of Hit and Run Cyberattacks

Security researchers at Kaspersky Lab say they have uncovered details related to a series of “hit and run” attacks against very specific targets. In a blog post on Kaspersky’s Securelist blog, the researchers said, “We believe this is a relatively small group of attackers that are going after the supply chain—targeting government institutions, military contractors, maritime and ship-building groups, telecom operators, satellite operators, industrial and high technology companies and mass media, mainly in South Korea and Japan.”

What’s most unique about the data theft campaign, which Kaspersky calls “Icefog,” is that after the attackers get what they want, they don’t hang around, using the backdoors installed on the victims’ networks to continue exfiltrating data. They go in knowing exactly what they’re after, take only the target information, then sweep up, turn off the lights, and close the door behind them.

Kaspersky Lab said that it has observed more than 4000 unique infected IPs and hundreds of victims. Some of the companies targeted during the operation, which began in 2011, include defense industry contractors Lig Nex1 and Selectron Industrial Company, shipbuilders such as DSME Tech, and Hanjin Heavy Industries, telecommunications firms such as Korea Telecom, and even the Japanese House of Representatives and the House of Councillors.

Kaspersky has since published a full report (pdf) with a detailed description of the backdoors and other malicious tools used in Icefog, along with a list of ways to tell whether your system has been compromised. The researchers have also put up an FAQ page.

iPhone Break-ins and Countermeasures

Someone tinkering with his Apple iPhone figured out a way to bypass its lock screen, the first line of security for the gadget other than keeping it in your pocket. This week, Apple released its latest countermeasure: an iOS 7 software update that fixes the security hole that allowed an unauthorized user to access information including the handset owner’s e-mail, Twitter, Facebook, and Flickr accounts.

According to Forbes' Andy Greenberg, “swiping upwards on the lockscreen to bring up the iOS Control Center, then opening the alarm clock app, then holding down the power button to show the ‘power off’ and ‘cancel’ options, then tapping ‘cancel,’ and finally quickly double-clicking the home button to bring up the multitasking screen for various apps,” made those apps accessible.

It’s amazing what people with loads of time on their hands eventually stumble upon.

That news came the same week it was revealed that someone had found an even more involved method for fooling the iPhone 5’s fingerprint sensor. According to Marc Rogers, a researcher at the mobile security firm Lookout, it’s possible but highly unlikely that you’ll be the victim of his hack, which he detailed in a blog post (“Why I Hacked Apple’s TouchID, And Still Think It Is Awesome.”). To give you an idea of just how remote the possibility of your phone being duped using his technique, here are a few of the steps Rogers mentions: “You take the cleaned print image and without inverting it, print it to transparency film. Next, you take the transparency film and use it to expose some thick copper clad photosensitive PCB board that’s commonly used in amateur electrical projects. After developing the image on the PCB using special chemicals, you put the PCB through a process called ‘etching’ which washes away all of the exposed copper leaving behind a fingerprint mold.”

In other words, you can rest easy.

Photo:Charles Dharapak/Associated Press

IT Hiccups of the Week: S&P Warns Stock Exchanges Shape Up Their IT or Else

Last week was much like the week before, with U.S. state governments still struggling to make their newly modernized unemployment insurance systems work correctly. We’ll get to them shortly, but we'll start off with a warning by the financial rating services company Standard & Poor’s (S&P) to the world’s stock exchanges over their continuing “technical snafus.”

S&P Tells Exchanges Downgrades in Store if Technical Problems Continue

In a warning shot across the bows of all the world’s stock exchanges, the rating services company S&P published a report last week with the very unsubtle title, “Exchanges' Technical Glitches Reveal Growing Operational Risk—And Could Trigger Downgrades.” The report outlined S&P’s concern that the continuing and seemingly increasing operational problems the exchanges are experiencing—including yet another one, almost on cue last week, at the Options Price Reporting Authority—would mean that the ratings agency might have to factor in a higher level of operational risk in its assessment of an exchange’s credit worthiness. That, in turn, could trigger a downgrade in an exchange’s credit rating, especially if it appears the exchange doesn’t maintain sufficient liquid capital to cover a major operational meltdown, including so-called black swan events.

The S&P report includes a nice table of 25 operational "snafus," as it likes to call them, that happened between March 2012 and August 2013. There have also been at least three more this month (including the aforementioned one at the Options Price Reporting Authority) that could have made the list. We have covered most of the technical problems here at the Risk Factor, including many that occurred before 2012, and were therefore not included in the list. Among these were the multiple technical problems that plagued the London Stock Exchange and the New York Stock Exchange.

The S&P warning came in the wake of the U.S. Security and Exchange Commission telling the exchanges basically the same thing in not so many words. The SEC gave the exchanges 60 days to come up with an approach that would increase their resilience and improve their management of operational risks that turn into operational problems or crises. The government didn't overtly threaten the exchanges, but it was definitely implied that penalties might be forthcoming if the exchanges don't get their act together.

The S&P concluded its report in this way: “We haven't lowered any ratings as a result of these issues so far, but exchanges whose operational risk exceeds industry averages could see rating actions down the road.”

Well, neither the exchanges nor their CIOs can say they weren’t given sufficient warning.

Obamacare Health Insurance Exchanges Experiencing Tech Issues

Next week,  on 1 October, the public health exchanges being created under the U.S. Patient Protection and Affordable Care Act (pdf) are slated to open. As I recently mentioned, concerns are being voiced over the security, or presumed lack thereof, of the federal data services hub as well as the state exchanges. In August, a Forbes article claimed that a major reason behind the security concerns is that, “The administration knows that if the exchanges don’t open on time, there will be a ton of bad press. So they are much more likely to attempt to launch the exchanges without adequate privacy safeguards.”

Whether IT security is lacking or not may be debatable (the Federal government claims that security hasn’t been comprised, at least at its end), the intense pressure to go live next Tuesday is not. For instance, April Todd-Malmlov, the head of Minnesota’s new health exchange, admitted last week to the MinnPost that, “The schedule that we have is essentially putting a five- to 10-year IT project within a two- to three-year timeframe.”

An 11 September risk assessment of the Minnesota Department of Human Services' efforts to implement the exchange stated that, “The schedule will remain [urgent] due to tight timelines with no slack… There is little to no room for vendors to miss code drop dates, and we are severely lacking in adequate testing time.” The assessment lists the project as “red” or “at risk” of not being able to go live next week, but Todd-Malmlov brushed that assessment aside, telling the MinnPost that, “The entirety of the project scope has always been red — at least one part of it has always been red… It’s an indicator for us as to where to focus efforts.” She added that, “I would guess that every state has that same ‘red’ on [their risk assessments].”

Iowa’s Gov.Terry Branstead recently said his state’s health insurance exchange would be open on-time, but added that state officials “are going to try our level best to try to make it work.” There is nothing like a strong vote of confidence to reassure everyone.

The Wall Street Journal added to Branstead’s vote of confidence in an article last week. The WSJ reported that “the government's software can't reliably determine how much people need to pay for coverage” in the health exchanges in the 36 states where the federal government is running all or part of the exchanges. It quoted one senior health-insurance executive as saying, “There's a blanket acknowledgment that rates are being calculated incorrectly… Our tech and operations people are very concerned about the problems they're seeing and the potential of them to stick around.”

Government officials, however, downplayed the issue, stating that, “We continue working with [insurers] and we are confident that on Oct. 1, consumers will see accurate premium costs, including tax credits.” They did admit to the WSJ that, “We may encounter some bumps when open enrollment begins,” but also added that, “we’ll solve them.”

You could almost see those government officials making that last statement while crossing their fingers on both hands.

Oregon, which is one of 14 states that are running their own health insurance exchanges, announced in early August that it was pushing back the 1 October initial enrollment date by two to four weeks in order to give it time to ensure that any issues with its “technology, customer service and other internal processes” could be “ironed out” before going live. That was a move I bet many state health exchanges wish they could also take, but can't politically.

As the health insurance executive stated earlier, I expect this story will not be going away anytime soon. Next Tuesday ought to be pretty interesting from both technical and political perspectives.

California’s Unemployment Insurance System Goes from Bad to Worse

I wrote last week about the problems that Nevada and California were having with their new unemployment insurance systems. While Nevada claims to have finally started to get a handle on the worst of its technical problems (although not everyone agrees), the problems in California have apparently worsened. California’s Employment Development Department (EDD) had originally stated that only 5000 Californians were having trouble receiving their unemployment checks on time because of problems with its new system. However, the EDD admitted to the Associated Press that the actual number—which it revealed early last week had climbed to 50 000, a staggering increase in itself—had reached 185 000 as of Friday. Some 80 000 residents, the EDD acknowledges, still haven’t received the unemployment checks due them.

In related news, California’s unemployment insurance system problems affected, for the second week in a row, the accuracy of the U.S. Department of Labor's report on the number of Americans receiving unemployment benefits. If this keeps up, all of September's unemployment numbers may have to be discarded. The EDD says it’s working hard to fix the problem and promises a major fix over the weekend. But even if the work is successful, it may be weeks before everything, including the federal employment data, is finally straightened out. The EDD did offer its apologies (again) for the delays, which did little to pay the bills of those desperately awaiting their unemployment checks.

Minnesota’s Department of Human Services, which, as I mentioned earlier, has its hands full dealing with the implementation of the state’s Obamacare health insurance exchange, has also had to deal with a multi-day outage of the system used by health care providers to file their Medicaid claims. It isn’t clear how many healthcare providers have been affected, but angry healthcare providers are probably the last thing the agency needs to be dealing with at the moment.

Of Other Interest…

NASA Says Deep Impact Probe Feared Lost After “Software Glitch”

Orbital Sciences' Cygnus Capsule Needs Software Update

Honda Recalls 405 000 Vehicles for Airbag Microprocessor Repair

Arizona Lottery Replaces Tickets After Lottery System Doesn’t Allow Numbers to be Drawn

New DFW Airport Parking System Wildly Overcharges Thousands of Customers

Several Airlines Hit by Navitaire Reservation System Outage

Photo: Scott Eells/Bloomberg/Getty Images

This Week in Cybercrime: How Secure is the Obamacare Data Service Hub?

On 1 October, the public health insurance exchanges being created under the U.S. Patient Protection and Affordable Care Act [pdf] are planned to open. Last week, the U.S. Department of Health and Human Services Centers for Medicare & Medicaid Services (CMS) said that the Data Services Hub used to determine a person's eligibility for government subsidies for this new healthcare program was “ready to go,” according to Reuters

The CMS stated that, “The completion of this testing confirms that the Hub complies with federal standards and that HHS and CMS have implemented the appropriate procedures and safeguards necessary for the Hub to operate securely on October 1.”

The CMS announcement was very good news indeed considering that: (a) The data hub “can access personal records from seven different agencies—the Internal Revenue Service, the Social Security Administration, the Department of Homeland Security, the Veterans Health Administration, the Department of Defense, the Office of Personnel Management, and the Peace Corps—in order to determine eligibility for exchange subsidies and mandate penalties,” according to Forbes. (b) The HHS Inspector general  had released a report [pdf] in August seriously questioning whether the security of the exchanges could be completed by the 1 October deadline, given that a previously estimated 51-day security review was completed in only 10-days. And (c) the exchanges are likely a priority target for hackers.

The good news didn’t last long, however.

Government Executive magazine ran a story this week that reported that “due to limited means, Health and Human Services Department internal watchdogs do not intend to examine key security designs they did not have a chance to assess during [their] recent audit.” Apparently, while security risks posed by the exchanges are an important concern, it was only one among many needing assessment that was competing for HHS “limited resources.” The higher priority risks—technical and political—were that the exchanges weren’t going to be officially “open for business" on 1 October, the HHS implied.

Republicans have grabbed onto the security doubts as a political gambit to postpone Obamacare, which they are threatening to do by other means as well. Their case was bolstered a bit this week when a Minnesota exchange employee accidentally “sent an e-mail file to an Apple Valley insurance broker’s office on Thursday that contained Social Security numbers, names, business addresses and other identifying information on more than 2400 insurance agents,” the Star Tribune reported. While small from a numbers standpoint, it was significant from a political perspective.

Politics aside, the Obama Administration better hope that not only is the Data Services Hub secure, but that all the state exchanges are secure as well. If a major breach occurs at an individual state exchange, the public will likely view all health exchanges regardless of ownership as being insecure. With “software issues” already occurring in exchanges (Iowa’s Gov.Terry Branstead this week basically said to expect problems with the state's exchange), any data breach could sour more of the public on Obamacare. (A recent survey showed that 53 percent of Americans view the law unfavorably.)

Security a Wee Bit Lax at NSA

NPR radio’s Morning Edition interviewed National Security Agency’s chief technology officer, Lonny Anderson and other unnamed government officials this week, who provided more details on how Edward Snowden was able to make off with the treasure trove of highly classified NSA documents without getting caught. According to the NPR interview, as part of his job, Snowden was able to access part of the NSA's intranet website where the documents he stole were put “so NSA analysts could read them online and discuss them. Anyone with the right top secret clearance could visit that page and read the documents. … As a systems administrator, Snowden actually had the responsibility to go to that intranet page and move especially sensitive documents to a more secure location.”

In fact, Snowden was “actually observed accessing secret documents, but the assumption was he was just doing his job.” It was, the officials admitted, the “prefect cover” for someone wanting to steal documents.

The officials refused to discuss how Snowden actually was able to download the documents and leave NSA premises with them undetected. The hypothesis is that Snowden took them out on a USB thumb drive. As of last June, some NSA computers still allowed access to USB thumb drives, a practice highly restricted in DoD since 2008 because of a major security breach.

 Anderson told NPR that the NSA finally has a good idea of what Snowden took, but that has taken four months of effort even with hints from Snowden himself. So much for NSA’s vaunted 100 percent audit capability. Probably more worrying to NSA officials is that someone else already preceded Snowden but hasn't ever gone public about it.

Budding Cybercriminals Go To School

Finally, a story at ComputerWorld reports that security company RSA has found that “a growing number of experienced hackers have begun offering structured hacking courses for crooks seeking to make a career in cybercrime.”

According to the ComputerWorld story, “The courses range from the basics of online fraud to advanced courses on online anonymity tools, botnets, cleaning up electronic evidence and dealing with law enforcement.” In addition, the course curriculum follows those found at major academic institutions.

Hackers pay about $75 per lecture, with lectures on using credit and debit cards fraudulently highly popular. Lectures are usually held via Skype. The only drawback is that many of the courses are taught in Russian.

No, the story did not discuss the possible transferability of course credit towards a college degree.

Of Other Interest ….

Microsoft Issues Emergency Explorer Fix

Hackers Pool Efforts to Crack iPhone Fingerprint Reader

US Comptroller of the Currency Warns Banks of Cyberattacks

Sophisticated Cybercrime Groups Operating From China

Eight Arrested in U.K. Over Theft of 1.3 Million Pounds from Barclays Branch Computer System

“Snowden Effect” is Hurting U.S. Cloud Providers

Brazil’s President Postpones Trip to Protest NSA Spying

Cybercriminals Flock to Brazil

Brazilian Hacktivists Mistake NASA for NSA

RSA Warns Against Using NSA Breakable Security Algorithm

Photo: Stephen Lam/Getty Images

IT Hiccups of the Week: Computer Issues Create Misleading U.S. Jobless Numbers

Last week provided a nice variety of IT-related miscalculations, ooftas and other surprises. We start off this week with a follow-on to a story from last week that has created some unexpected consequences.

Computer Issues Create US Jobless Claim Number “Anomaly”

A little bit of background first. You may recall from last week's edition that at the end of August, Nevada’s Department of Employment, Training and Rehabilitation (DETR) took down its 30-year-old unemployment insurance system and began the roll-out of its new US $45 million UInv system, which has had a less than auspicious start including being offline for longer than predicted.

Nevada wasn’t the only state busily upgrading its unemployment system or having problems as a result. California’s Employment Development Department (EDD) has spent US $157.8 million upgrading the state’s 30-year-old unemployment payment processing system. California has the largest unemployment system in the U.S., disbursing some $33 million a day in unemployment checks.

According to an EDD statement over the weekend, there have been “some processing delays in [the] transition” to the new system, which began, like Nevada’s, at the end of August. The statement says that EDD has been “working around the clock to catch up on unemployment claims.” A news story quoted an EDD spokesperson reporting that about 5 percent of the claims—or roughly 20 000—were affected by the upgrade issues. She added that, “We apologize for the inconvenience to those affected.”

Now, for the punch line: Last week, the U.S. Department of Labor released its report on the number of Americans receiving unemployment benefits. The agency reported a drop of 31 000 from the previous week, for a seasonally adjusted total of 292 000—the lowest number since April 2006. According to the Wall Street Journal, drops like that are rare.

Fantastic news, eh, since it could indicate the end of the Great Recession! When the report first came out, economists were overjoyed, until they found out the numbers were “flawed;” then they were a bit upset, to state it mildly.

What the Labor Department's jobless report didn’t disclose was that the drop was related to the computer problems in Nevada and in California. Only after the report was release did the Department admit the discrepancy  to reporters. On top of that, it refused to identify the two states involved, even though it isn’t a mystery to reporters which two states were having troubles upgrading their unemployment insurance systems and reporting timely unemployment insurance claims numbers. In addition, the total claims number was also likely skewed downward because of the shortened work week following the U.S. Labor Day holiday. Finally, Massachusetts’ new unemployment system is said to be “riddled with problems” as well; it is unclear how those problems have affected the total claims number, either.

A Labor Department spokesperson defended the faulty report to the New York Times, saying, “When we get data, we have an obligation to put it out there.” And anyway, he emphasized, "the department did not recommend reading too much into any one week’s figure, at any rate.”

Of course, if the unemployment number is a positive one (and also accurate), the Labor Department doesn’t seemed constrained from touting it from the rooftops, or trying to bury it if is negative. Watch what happens when Nevada and California send in revised unemployment claim numbers later this month.

Stephen Stanley, chief economist at Pierpont Securities, said that the episode was a classic example of “bureaucratic ineptitude.” That is being way too nice.

Later this month, after spending US $69 million, Michigan is going to upgrade its 30-year old unemployment system. It may be déjà vu all over again in October.

United Airlines and Jet Blue Have System Problems

Last Friday, both United Airlines and Jet Blue experienced problems with their reservation systems. According to a story at Fox News story, Jet Blue reported that morning that an “IT system outage” caused by “connectivity issues” (supposedly at Verizon) had caused delays for about 60 flights. However, the Fox story also reported that, “flight tracker FlightAware said it recorded 70 flights that were delayed for more than an hour. Two hundred twenty two flights were delayed for more than 15 minutes.”

A horror more appropriate to the Friday the 13th date on the calendar was the disturbance that hit United Airlines. A computer error on its website slashed United ticket prices to zero (but with the US $2.50 security charge per leg intact). It took two hours before United realized what was going on and shut down its website. But by then, word of the pricing error had spread across social media, with lots of people announcing their good luck.

United, after thinking about it, decided it would, like Italian airline Alitalia did last year, honor the tickets. Apparently, the good publicity was thought to outweigh the costs, especially in light of the bad feelings United’s other recent computer problems have left with many of its customers.

Other carriers, like British Airlines, have not been so generous.

UnitedHealth Recalls EHR Software

Last week, UnitedHealth Group issued a recall of its software that manages electronic health records used in 35 hospital emergency facilities in 22 states. According to Reuters, UnitedHealth—the largest U.S. health insurer—found that an error in its Picis ED PulseCheck software caused some doctors’ notes on patient prescriptions to vanish.

A story at Bloomberg News says that UnitedHealth acquired the maker of the software, Picis, Inc. of Wakefield, Massachusetts, in 2010. Bloomberg also noted that this is the sixth recall of Picis electronic health record-related software since 2009.

In related EHR safety news, the Pennsylvania Patient Safety Authority issued an advisory last week to hospitals and other health care providers telling them to check the default settings in their EHR and computerized physician order entry (CPOE) systems. The Authority found 324 adverse patient events traceable to supplier-set system default settings not being reset to more appropriate ones matching the operating context of the healthcare provider.

I don’t doubt the suppliers have warnings about checking the defaults in their manuals, but hey, who reads a manual anymore?

Of Other Interest…

UK TSB Bank Website Crashes on Relaunch

University of Minnesota Students Find Extra US $1000 Billed

Amazon Cloud Disrupts Some Websites

Third Outage of New Jersey State Computer Systems Since July

“Technical Glitch” Affects Access to Thousands of Xbox Games

New York City Penalizes Two Testing Companies Over Exam Errors

Illustration: iStockPhoto

This Week in Cybercrime: Companies to FTC: Your Data Security Reach Exceeds Your Grasp

The U.S. Federal Trade Commission is wrong to claim broad authority to seek sanctions against companies for data breaches when it has no clearly defined data security standards, said panelists at a forum sponsored by Tech Freedom, a Washington, D.C., think tank that regularly rails against government regulation.

The event, held on Thursday, coalesced around the fact that in the last decade, the FTC has settled nearly four dozen cases after filing complaints based on its reasoning that a failure to have sufficient data security constitutes an unfair or deceptive trade practice. Two pending court cases, says a Tech Freedom statement, "may finally allow the courts to rule on the legal validity of what the FTC calls its 'common law of settlements.'"

Read More

IT Hiccups of the Week: A Bad Week for U.S. State Government IT

It’s been another relatively normal week in the land of IT inconveniences except perhaps for government computing systems, which is where we will concentrate our focus this week. We start off with the problems occurring in several U.S. states that recently introduced new IT infrastructure that is proving balkier than hoped.

Nevada, Massachusetts, and North Carolina Each Have Buggy New IT Systems

On 26 August, Nevada’s Department of Employment, Training and Rehabilitation (DETR) took down its 30-year-old unemployment insurance system and began the rollout of its new $45 million UInv system. In a press release issued that day, DETR announced (pdf) that the new system would be operational on 1 September; 51 000 or so Nevadans looking to file unemployment claims would have to wait until then.

However, the DETR enthused that after four years of development the new system would be worth the wait since it would, “allow claimants to view up-to-date information related to their individual claims. It will also give claimants access to their payment history and allow them instant, real time feedback on their unemployment claim.”

Unfortunately, the UInv system wasn’t ready for prime time until 4 September. The DETR "explained" (pdf) the delay in a subsequent press release, citing a number of undisclosed “minimal issues.” The DETR release went on to say that the department was “being very conservative” with the launch of the new system, and it asked for “patience as we gradually ramp up the new system to full deployment over the coming days.”

The Nevadans affected by the delay were clearly not amused.  The DETR had promised that when the new system was up and running, “claims and benefit services will continue as normal.” But by the 4th, it was backpedaling on that promise. The ballyhooed new online system wasn’t working and wouldn’t be for another three days. And according to the Las Vegas Sun, Nevadans couldn’t reach anyone at the DETR to get help with their claims. The reason was a study in irony: the phone lines were unexpectedly overwhelmed with callers when the DETR encouraged claimants to call in because online access had not yet become operational.

A DETR spokesperson said unhelpfully, “keep calling, relax and we will get to you.”  After hours upon hours of waiting on hold, many Nevadans gave up. Governor Brian Sandoval is said to be aware of the problem, but there is little that he can do in reality.

DETR says no one will miss out on their unemployment checks since the claims would be backdated, but also admitted that it may still take a while for the payments to be made.  Hope none of those unemployed Nevadans have any pressing bills to pay. Future upgrades to the UInv system are schedule for later this year and early next, which everyone no doubt hopes will be smoother.

Massachusetts residents who receive unemployment insurance are also unhappy with that state’s new unemployment benefits computer system that was launched in July. The $46 million system ($6 million over budget) has been plagued by problems and is “unable to make proper payments to hundreds of financially strapped workers hunting for jobs,” according to a Boston Globe story.

The system's contractor, Deloitte Consulting, has until the end of the month to fix the system without penalty, the Globe reports, but the newspaper also states that, “It's unclear what remedies are available to the state if the system is still not working properly after that.”

Deloitte says it has hired extra workers to help with the backlog of unemployment claims as well as notices sent out by the new system to unemployed workers demanding repayment for money they never received.

Former secretary of Labor and Workforce Development Suzanne Bump, who is now the state auditor, listed the upgrading of the new system as one of her accomplishments during her tenure at the labor department, but the Globe states that she is now trying to disassociate herself from the project as quickly as possible. A big surprise, eh?

Michelle Amante, the state official now in charge of the project, used to work for Deloitte on the project. She claims that despite all of the problems, “we fundamentally believe that the system is working.” Another big surprise.

Finally, joining (or remaining in) the ranks of unhappy constituents this week are North Carolina residents and businesses. The state recently rolled out two new systems, one called NCFast and the other NCTracks. NCFast (North Carolina Families Accessing Services through Technology), which was “soft-launched” in mid-summer (the system will not be finished until 2017), is the new N.C. Department of Health and Human Services computer system that is supposed to streamline the work activities and business processes of the department and county social services agencies so that more time can be spent helping those requiring public assistance and less on bureaucratic tasks

However, there have been ongoing issues with the $48 million system that have caused many families on food-assistance to go without their benefits. The state is blaming the counties for the problems, while the counties are blaming the state.

The same department has another headache in the form of its new NCTracks system. (I have no idea what that is an acronym for, if in fact it is one.) On 1 July, the department launched its controversial $484 million system in the wake of a state audit (pdf) released in May that cast doubt on whether the system—which was $200 million over budget and two years late—was ready to go live. The audit cited, among other things, the lack of testing of key system elements.  

The Department of Human Services insisted on 1 July that there was nothing major to worry about, regardless of what the audit reported. It conceded that there might be an “initial rough patch of 30 to 90 days as providers get used to using the new system,” but that there should be smooth sailing after that. Well, it has been a very rough patch indeed for many providers, who are, after 70 days and counting, still very unhappy with the system. The department has even had to mail emergency paper checks to over a thousand providers who couldn’t get their claims accepted by the new system and were facing financial hardship.   A Triangle Business Journal story from last week reported that the NCTracks “has missed its own targets nearly across the board, some by significant amounts.

With the ongoing problems at both NCFast and NCTracks, North Carolina lawmakers are now going to get involved. Exactly how they intend to improve the situation is a bit of a mystery.

Australian Very Happy over Credit Card Glitch

According to a story from Australia News Nine, Carlo Spina wanted to buy a magazine at a BP service station in Sydney. However, Spina discovered that he did not have enough cash on him, which meant he had to pay by credit card. When he tried to do so, Spina had trouble getting the card to work. The few seconds spent fussing with the card reader probably saved Spina's life as an out of control SUV crashed into the service station in the path he would have taken to leave.

You can watch the very close call via a Youtube video here. Spina wasn’t hurt, but was shaken up. I don't know if he ever did read his magazine.

Volvo Recalls 2014 Models

Finally, Volvo announced that it was recalling 8000 of its 2014 model year cars in the United States and Canada because of “a software glitch that could drain the battery and cause headlights, windshield wipers and turn signals to malfunction.” The models affected include S60 and S80 sedans and XC60 and XC70 crossovers.

Of Other Interest…

Tesco Pricing Error Sells White Chocolate Oranges for 1p

DC Metro Computer Problems Resolved

New Zealand Visa Bungle Blamed on Overloaded Computers

New School Computer Software Causing Confusion in Seattle

London Black Cab Production Restarts

NASA’s LADEE Glitch Fixed

Photo: iStockphoto

This Week in Cybercrime: Middle East’s Upheaval Breeds Hacktivists

Unrest in the streets of Egypt and Syria has led to thousands of deaths and a lack of personal security in many public spaces. But the civil wars raging there are turning out to be the backdrop for diminished security across online networks. McAfee, the cybersecurity firm best known for its online antivirus solution, told Reuters this week that more than half of the cybercrime activity now occurring in the Middle East can be characterized as “hacktivism” by politically motivated programmers looking to sabotage opposition institutions or groups.

“It’s difficult for people to protest in the street in the Middle East and so defacing websites and [carrying out] denial of service (DOS) attacks are a way to protest instead,” Christiaan Beek, director of incident response forensics for McAfee in the Europe, Middle East and Africa (EMEA) region, told Reuters.

The targets have overwhelmingly been entities linked to the region’s economic underpinnings, which in most cases means crude oil. Cyber attacks in the region are reportedly focused on Saudi Arabia, the world’s leading oil exporter; Qatar, the top supplier of liquefied natural gas; and Dubai, the region’s aviation, commercial and financial hub.

Gert-Jan Schenk, McAfee president for the EMEA region, told Reuters, “Ten years ago, it was all about trying to infect as many people as possible. Today we see more and more attacks being focused on very small groups of people. Sometimes malware is developed for a specific department in a specific company.”

Read More

Are STEM Workers Overpaid?

One of the strongest reasons given by those trying to entice more students to enter the STEM education pipeline is the “earnings premium” STEM workers make in comparison to non-STEM workers. Typical is the statement by the U.S. Department of Commerce press release from 2011 that, “STEM workers command higher wages, earning 26 percent more than their non-STEM counterparts.”

Further, the Commerce Department press release quotes U.S. Secretary of Education Arne Duncan’s plea to prospective STEM students that, “A STEM education is a pathway to prosperity – not just for you as an individual but for America as a whole. We need you in our classrooms, labs and key government agencies to help solve our biggest challenges.”

However, not everyone is happy as Duncan with STEM workers earning a premium for solving those big challenges, instead believing that the U.S. would be even more competitive and have a more equitable society if the earnings premium disappeared. For instance, former Federal Reserve Bank Chairman Alan Greenspan, speaking at a U.S. Treasury conference on U.S. Capital Markets Competitiveness put it bluntly: “Our skilled wages are higher than anywhere in the world. If we open up a significant window for skilled [guest] workers, that would suppress the skilled-wage level and end the concentration of income.”

Greenspan, to ensure everyone got the point, added, “Significantly opening up immigration to skilled workers solves two problems. The companies could hire the educated workers they need. And those workers would compete with high-income people, driving more income equality.”

[For a detailed examination of how effective this policy could be, I strongly suggest you read Eric Weinstein’s National Bureau of Economic Research draft working paper titled, “How and Why Government, Universities, and Industry Create Domestic Labor Shortages of Scientists and High-Tech Workers,” on the active suppression of STEM Ph.D. salaries by way of false National Science Foundation claims of a STEM shortage coupled with aggressive lobbying efforts to change STEM guestworker policies in the late 1980s to early 1990s. While the NSF eventually apologized for its misrepresentations to Congress in 1992 and admitted that there was in fact a surplus of STEM workers, the damage was already done with the fallout continuing into today.]

Greenspan is not alone in his thinking that STEM worker salaries should look like a lot more like non-STEM worker salaries. In March, over “100 executives from the technology sector and leading innovation advocacy organizations” sent an open letter to President Obama begging that he and Congress would approve the expansion of H-1B visa program beyond the 85 000 visa limit today (including 20 000 reserved for foreign graduates with advanced degrees at U.S. universities) to a minimum of 115 000 per year and possibly as high as 300 000 within a decade (not to mention the granting of permanent legal status to an unlimited number of foreign students who earn graduate degrees from U.S. universities in STEM subjects).

The  executives, whose companies have spent millions of dollar lobbying on the issue, collectively wrote in their letter that, “One of the biggest economic challenges facing our nation is the need for more qualified, highly‐skilled professionals, domestic and foreign, who can create jobs and immediately contribute to and improve our economy.”

Four company executives from IBM, Intel, Microsoft and Oracle claimed in the letter they had 10 000 openings that they apparently could only fill with guestworkers. That was, of course, before IBM announced its layoff, or Intel acknowledging that it was going to slow down its hiring as well as cut production at some of its U.S. plants which might lead to workers being let go. It will be interesting to see whether Microsoft CEO Steve Ballmer’s successor decides that the company payroll needs a bit of trimming, like Lou Gertsner did at IBM. And one other company on the letter list is Cisco, which also announced layoffs since the March letter. Maybe the companies can just trade workers.

While the tech company executives insist that their motive is not to cut payroll costs by increasing the supply of guestworker labor, few STEM workers believe their claims that there exists a “technology skills gap” that only guestworkers can fill, any more than there is a U.S. manufacturing skills gap. For instance, various surveys have claimed there is a skilled manufacturing work force shortage at between 300 000 and 600 000, for example.

However, others digging through the veracity of those claims like the Boston Consulting Group point out that the manufacturing skills gap is in reality closer to 80 000 and 100 000, and would be even less if employers were to increase their pay or hire less-skilled workers and train them, both of which employers seem highly reluctant to do. As the BCG study states, “Trying to hire high-skilled workers at rock-bottom rates is not a skills gap.”

Likewise, tech companies are definitely not interested in paying higher wages, something they were unhappily forced to do during the dot-com boom period (and which some tried to avoid afterwards as well by tacitly agreeing not to poach employees from one another until they got caught). The fact that thousands of these unfilled jobs stay unfilled for long periods of time also seems an indication that they are not nearly as critical or valuable as the companies make them out to be. Netflix, for example, pays its 700 software engineers 10 to 20 percent above its competitors in Silicon Valley, allowing it to “hire just about any engineer it wants.”  As Netflix clearly demonstrates, an expanding U.S. company in a competitive marketplace that needs technology workers can get the U.S. technology workers it wants, if it really wants them.

Anuj Srivas, the technology and business reporter of the Indian paper The Hindu, pointed out earlier this year that despite all the rhetoric, the “H-1B visa farce” as he calls it is indeed “about the profit margins of Indian and American IT companies,” something that Vivek Wadhwa, an academic and entrepreneur who advocates more H-1B visas, has acknowledged. Wadhwa candidly wrote in 2008 that, “I know from my experience as a tech CEO that H-1Bs are cheaper than domestic hires. Technically, these workers are supposed to be paid a ‘prevailing wage,’ but this mechanism is riddled with loopholes. In the tech world, salaries vary widely based on skill and competence. Yet the prevailing wage concept works on average salaries, so you can hire a superstar for the cost of an average worker. Add to this the inability of an H-1B employee to jump ship and you have a strong incentive to hire workers on these visas.”

Exactly how many “superstar” H1B guestworkers are working in the U.S. is in some doubt. For example, currently the majority of H1B guestworkers are Indian (pdf). Yet, Indian companies themselves complain over the cost of making the vast majority of engineering graduates employable (pdf).  Back in 2006, when India graduated some 350 000 engineers, employers there estimated that about 10 percent to at best 25 percent were employable by multi-national companies.  In 2012, however, there were over 850 000 graduates from Indian engineering colleges. It is doubtful that the quality of engineering education in India has been maintained to even 2006 levels given the huge increase in graduates and engineering schools.

Given the long standing complaints by tech company executives of a skill shortage and the need for more guestworker labor (e.g., in 1983, John Calhoun, the director of business development at Intel testified to the U.S. Congress in regard to the need for more technology worker immigration, “The problem is absolutely one of a shortage and not one of lower-cost labor. We in the industry have been forced to hire guestworkers in order to grow.”), it is interesting to see how engineering and computer professional salaries have risen. The reason I say risen is that rising wages is usually considered the best indications of a shortage, according to the RAND Corporation.

Using data from a Northwestern University 1995 study (pdf) that provided starting engineering graduate salary information from 1950 to 1994 and normalizing it to 2011 U.S. dollars, you see that average starting salaries peaked around 1970 at $61 200 and then dropped slowly but surely to $52 470 in 1995. The engineering class of 2011 average starting salary (pdf) was $59 590, thanks in part to the dot.com demand, which pushed starting salaries up about a decade ago.

This is shown in the IEEE-USA published salary survey data. In 1972, the median salary was $101 300 in 2011 dollars, whereas in 1975, the median salary had dropped to $98 460.  By 1992, median salaries had fallen to $91 823. Median salary recovered during the mid to late 1990s, and hit a peak of $122 315 in 2002. The 2011 IEEE survey data shows the median salary to be $115 790.

For computer professionals, the last ten years have also shown little change in salary as well. Compiling a decade’s worth of published DICE salary data, the average salary in 2011 constant dollars was $86 823 in 2001, and was only $83 858 last year. According to a story from CNN at the time, a software developer’s average salary in 1990 ranged from $84 750 to $101 600 in 2011 dollars.

It is hard to see from the salary data that there has been a huge spike in engineering salary over the last thirty plus years—the same period that tech executives have been claiming that they have been faced with a tech skill shortage, the same period during which numerous tech companies apparently succeeded and made pretty decent profits. They may have had somewhat of a case to whine about in the mid to late 1990s in regard to engineering jobs, but the study by RAND found little evidence of one even then. The salary data definitely doesn’t indicate a STEM shortage is occurring now.

As an EPI analysis in April which found no STEM worker shortage in the US noted, “policies that expand the supply of guestworkers will discourage U.S. students from going into STEM, and into IT in particular.”

If that happens, and then foreign STEM workers decide to stay home because the salaries they are also earning wages in the U.S. approach those of non-STEM workers, U.S. high tech executives and the government will have no one to blame but themselves.

Photo: Muharrem Oner/Getty Images

Advertisement

Risk Factor

IEEE Spectrum's risk analysis blog, featuring daily news, updates and analysis on computing and IT projects, software and systems failures, successes and innovations, security threats, and more.

Contributor
Willie D. Jones
 
Load More