The State's Health Czar is Blaming Accenture for a program he was told would likely fail
In a misguided, poorly executed effort to let the private sector bring “efficiency” to Texas government, the state has squandered at least $100 million, cheated hundreds of thousands of needy Texans out of benefits, and now risks millions of dollars in federal fines for botching things up so badly.
This is not the promised land of free-market bounty that Republican lawmakers and right-wing think tanks described in 2003, when they convinced the Legislature to enact one of the most ambitious privatization plans any state has ever attempted. Back then, the talk was of how wonderful life would be as projections of $600 million in savings danced a lively jig through committee rooms and onto the House and Senate floors.
Now that the music has stopped, state officials are risking serious physical injury rushing to blame the debacle on Accenture Ltd., the private consulting giant that won the $899 million contract to set up call centers to help Texans get government benefits like food stamps, welfare, and Medicaid.
But many of those same officials—particularly Albert Hawkins, head of the state’s Health and Human Services Commission—were warned repeatedly that the program was being rushed, was never properly tested, and risked failure. The project plunged ahead anyway. An Observer review of thousands of pages of state documents reveals that while Accenture certainly made mistakes, it was Hawkins who relentlessly pushed the program toward launch on January 20, 2006—what HHSC described as the “go-live” date—despite numerous warnings from federal officials, Accenture, and his own auditor that the network of call centers wasn’t ready. HHSC officials also intentionally scaled back tests they knew the system would fail. And Hawkins moved the call-center project forward at two separate stages without obtaining prior federal approval—despite repeated warnings from the feds—actions that cost the state millions in federal funds.
“There were red flags,” says state Rep. Abel Herrero, a Corpus Christi Democrat who’s leading the legislative committee investigating the Accenture debacle. “There’s deep concern why the agency would proceed with an aggressive schedule with a program that was known to be deficient.”
The consequences of HHSC’s mismanagement are clear. The anticipated five-year savings of more than $600 million from the call-center plan haven’t materialized. In fact, the Accenture contract ran over budget. The state has already spent $30 million cleaning up the mess.
Moreover, thousands of eligible families mistakenly were dropped from government programs such as Medicaid and the Children’s Health Insurance Program—gaffes that cost the state tens of millions in federal matching funds. HHSC’s inability to enroll eligible families for benefits risks further fines and sanctions of up to $30 million from the federal government.
As head of the state’s five health and human service agencies, Hawkins is perhaps Texas’ most powerful unelected official. He remained a driving force behind the call-center plan even after former state Rep. Arlene Wohlgemuth, the Burleson Republican who drafted it, left state government to run unsuccessfully for Congress. His hunger to enact call centers has puzzled some observers in the Capitol. Appointed by Gov. Rick Perry in early 2003, Hawkins is a former budget director under then-Gov. George W. Bush. He’s an ace numbers man respected by lawmakers of both parties. While conservative politically, he’s not known as a rigid ideologue. All of which makes his mishandling of the call-center implementation peculiar. (Hawkins declined an interview request for this story.)
Perry has reappointed Hawkins to a new term. But the state Senate has thus far withheld his confirmation. It remains to be seen whether he’ll stay in his post, though there is no apparent attempt to oust him over the costly fiasco.
It began as a grand experiment to graft the principles of the free market and the private sector onto Texas’ system for screening and enrolling poor families in government benefit programs. In the past, those wanting to sign up for Medicaid or food stamps or welfare had to visit the nearest state office. A state employee would assess their needs, verify income, and proffer the proper forms for signing.
Under the new model, the state planned to close dozens of local offices, fire thousands of state workers, and replace them with four call centers run by a private company. Texans would access the social safety net much the same way they complain to the cable company about poor service. No state had ever outsourced such a core government function. But Republican lawmakers pitched their plan as the vanguard of a new movement, and in June 2005, Texas inked the largest contract in state history, awarding Accenture $899 million to run the show.
It never worked, and just 20 months later the great experiment was in ruins. After four months of dismal reports on Accenture’s continuing inability to make its call centers function, the HHSC tore up the contract on March 13. In the days after the massive deal came to an inglorious end, HHSC officials made clear in numerous statements that they put the blame on Accenture. Hawkins told a legislative committee that Accenture had assured him the call-center system was ready.
He was not telling the whole truth and nothing but, the Observer document review shows. Hawkins and HHSC were, in fact, warned that the system was likely not ready. On two separate occasions, Hawkins failed to get federal approval for the plan despite repeated warnings from federal officials that prior authorization was needed to access federal money. Because Texas was attempting a radical reform that no state had ever tried, the feds had a keen interest: Several federal agencies would have to give their blessings at every major stage, or Texas would risk losing federal funds. This was particularly true of the food stamp program, which has slightly more stringent enrollment requirements.
The federal Food and Nutrition Service warned HHSC, according to memos, that Washington needed to approve the call-center plan before HHSC put out a call for private bidders. Yet HHSC published its request for proposals without federal approval on July 22, 2004. Four days later, the nutrition service wrote HHSC, “Please be advised that, pending our approval of your RFP, you are proceeding at your own risk and that no federal funding for this initiative can be provided … until we have fully approved it.” FNS eventually approved the funds.
Trouble arose again in early 2005 as HHSC began negotiations with Accenture. FNS warned Hawkins that it had to approve the final contract before it was signed and that FNS needed 60 days to do that. Despite early and repeated warnings, HHSC gave the feds a lead time of exactly one day. The state submitted the Accenture contract for federal review on June 28, 2005, according to agency documents. On June 29, 2005, HHSC signed its deal with Accenture. The feds seemed baffled. “If Texas had followed the normal course of events, the State would have secured the approvals. … Since the contract was signed before obtaining our approval, the funds that are expended now must be State funds and may not be reimbursed,” FNS wrote on July 14, 2005.
FNS did approve the contract in November 2005. But in the intervening five months, HHSC lost about $12 million in federal funds, according to an agency spokesperson. (FNS so far has denied the state’s attempts to recoup that money.)
On March 7 of this year, Herrero, who heads the subcommittee of the House Committee on Human Services that is investigating the Accenture fiasco, asked Hawkins at a hearing why his agency brushed off the feds. “We still believe that all those elements do not require federal approval,” Hawkins said. “But that aside … we have been in constant communication with the FNS. It’s not that they were unaware of the elements of the contract or how the state was moving forward. They had not finalized their approval on it.” He later claimed that the Accenture contract was the only part of the call-center project that didn’t gain prior federal approval. “The one that did not was the contract—not prior approval. There was for the RFP,” Hawkins said. That statement isn’t accurate, according to a review of FNS memos.
When the feds finally approved the Accenture contract, they did so with conditions: FNS granted federal funding for only the project’s first three months. The agency was concerned that HHSC was rushing out a system that would deny benefits to eligible families—a concern that would prove prophetic. “We continue to have concerns about the seemingly unrealistic time frames for implementation of the state’s plan,” the feds wrote to HHSC.
Indeed, the state was in a major rush. After signing the contract, HHSC gave Accenture just seven months to set up its call centers before the January 20, 2006, “go-live” deadline, when the centers would begin fielding calls from needy Texans in two Central Texas counties. HHSC wanted the call centers to handle all enrollment statewide by December 2006. Could HHSC and Accenture really transition all enrollment and eligibility for complex programs that serve millions of Texans in 11 months? The feds didn’t think so. “[FNS] had some real anxiety about the aggressive conversion schedule that the state had developed,” Hawkins recently conceded to Herrero’s subcommittee. “[FNS] did advise us that it should be slower.” He ignored their advice.
As the go-live date fast approached, HHSC received several more warnings of flaws in the system and admonitions that its rollout schedule was too ambitious. Two independent verification audits, one commissioned by HHSC and one by FNS, reported serious problems with the call-center system, according to a report issued last fall by the Texas Comptroller’s office. HHSC’s own inspector general’s office also warned Hawkins of flaws in the call center’s computer system. In late December 2005, Accenture warned HHSC of “red,” or critical, system failures. The company, formerly a division of Arthur Andersen LLP, a defunct accounting firm, had run into a serious technology problem.
An Accenture subcontractor planned to use computer software to record applicant information at call centers. Problem was, the contractor’s software couldn’t interface with the state’s computer system. The inability of the two systems to talk to each other would prove a critical obstacle. To make matters worse, the state software was new. HHSC was rolling out a new computer system, known as Texas Integrated Eligibility Redesign System (TIERS—pronounced “tears”). The complex system is aptly named: It has been a constant source of heartache for state officials since its botched creation by accounting firm Deloitte and Touche LLP. Texas has spent the past five years and tens of millions trying to fix TIERS, with intermittent success. It has been perhaps an even larger disaster than the call-center contract. Yet HHSC had decided to package the call-center system with TIERS and planned to roll them out together.
HHSC officials say Accenture promised them it could “work around” the software incompatibility problem. Accenture would hire more workers, at the company’s expense, to do “dual entry,” meaning tediously typing in the same information twice—once into each software program. It was an ironic solution since the call centers were supposed to make the whole process more efficient. The “work around” resulted in numerous bureaucratic errors. Anne Heiligenstein, a deputy commissioner at HHSC, recently told lawmakers that, “We went to pilot [stage] based on the … information the vendor brought forward that there was a viable solution [to the technology problems]. We accepted that. They offered that, I’m sure, in good faith. It did not achieve what it needed to achieve.”
Taking Accenture’s assurances at face value, HHSC decided not to test how well the call centers’ technology functioned before the rollout. In the rush to launch the system, HHSC officials instructed Accenture to scale back several elements of its final readiness testing.
Accenture’s report on readiness testing in January 2006 reads, “Due to significant delays in testing for the rollout coupled with a decision to go live on January 20, 2006 without regard to actual results of the Readiness Assessment Tests, [Accenture] and HHSC agreed to a scaled back Readiness Assessment Test that would only test critical technical components.” In fact, components of the system that would later prove to be major flaws—staff levels, staff training, proper routing of calls, fraud prevention, and computer compatibility—were intentionally dropped from the readiness tests, according to Accenture documents. The software also wasn’t tested because HHSC officials already knew the computer systems couldn’t talk to each other and would fail the test, according to the HHSC. Agency officials say a test wasn’t needed because they believed Accenture’s promises that manual entry would solve the problem.
What about the other flaws omitted from the readiness tests? HHSC spokesperson Stephanie Goodman says that Accenture’s “Readiness Assessment Tests” were just one form of testing. Whatever elements Accenture left out, she says, HHSC tested on its own before the rollout.
Still, problems were evident. Herrero says the call-center launch should have been delayed. “The standards and quality-assurance levels were lowered to meet the schedule set by the commission,” he says.
Despite the warnings, HHSC left itself without a backup plan. In late 2006, the agency notified nearly 3,000 state enrollment and eligibility workers that they would soon lose their jobs. Predictably, this led to an exodus. More than 800 veteran enrollment workers took jobs elsewhere in state government, and hundreds more simply walked out. The loss of institutional memory proved disastrous when the call-center plan failed, and HHSC found it needed the state workers, some of whom had done enrollment for decades. “What’s troubling to me is that the commission would encourage employees to go elsewhere before there was a successful, integrated eligibility program,” Herrero says. “There were problems. The go-live date was fast approaching, and despite that, [state] employees were advised that their jobs would no longer be needed.”
As the warnings had predicted, the call-center system became a mess. The failures are well documented. Once Accenture took over screening in two Central Texas counties, and screening for the Children’s Health Insurance Program statewide, enrollment in all benefit programs began to drop.
Enrollment in CHIP declined by as much as 27,000 kids in early 2006. (It’s difficult to determine how much of the CHIP shrinkage was caused by the call-center failure; the program had been declining for several years before the Accenture deal.) The children’s Medicaid program had expanded for years prior to the call-center debacle. After the go-live date and the loss of state workers, 118,000 children lost Medicaid coverage in just nine months. Denying these CHIP and Medicaid benefits deprived the state of at least $50 million in federal matching funds.
In the end, the call centers were overwhelmed by the number of calls, inadequate and poorly trained staff, and malfunctioning technology (Accenture’s “dual entry” solution was far too mistake prone and inefficient). Wait times and backlogs of applications swelled. Accenture sent out renewal letters dated May 8 that informed clients their deadline for reapplying for benefits was April 15. One Accenture fax number mistakenly routed benefit applications to Seattle.
After four months, HHSC halted the call-center rollout on May 10, 2006. It never resumed. HHSC scrambled for a backup plan, rehiring thousands of state workers a
d paying out “rete
tion bonuses” to keep temporary screeners. Cleaning up the Accenture fiasco cost the state $30 million, according to HHSC, money it still hopes to recoup from Accenture.
For now, state employees once again are handling screening and enrollment. But with the loss of so many longtime enrollment workers, a system that functioned well before the call-center experiment has suffered permanent damage. Texas used to routinely earn good-performance bonuses from the food stamp program for its mistake-free enrollment system. In 2004, for instance, the feds paid the state $150 million in bonuses. Now it’s the opposite. For the first time in years, Texas’ food stamp enrollment error rate—the number of applications the state botches—exceeded the national average in 2006. If the error rate breaks the national average again this year, Texas will face penalties and sanctions from the federal government that could approach $30 million.
In a recent hearing, Herrero asked Hawkins why HHSC didn’t heed warnings from the federal government, auditors, and the agency’s own contractors. “I thought it better to follow the direction of the Legislature,” he said, “and to achieve the savings [from the call-center plan] … to fit within the [HHSC] budget.” In other words, he ignored the warnings so his agency could realize the estimated savings. It was an ironic answer since the call-center rollout ended up costing the state millions.