Mark Romano gripped the steering wheel and tried to keep his car from swerving into another commuter on the busy Illinois tollway.
Stress? It was December 2007, and Romano was a senior manager at Allstate and its top expert in Colossus, a program that calculates how much a person might be paid for an injury claim. He was in charge of two projects to “tune” and “recalibrate” Colossus, work he knew could affect payments to thousands of people.
Colossus was part of a quiet revolution in the insurance industry. Before the early 1990s, insurance was a decidedly human endeavor, especially when it came to setting rates and paying claims. To set premiums, insurers relied on computations from their actuaries — mathematical wizards armed with statistics and tables that assess various risks. When it came to paying claims, insurers often sent adjusters into the field, where they met face-to-face with people injured in car wrecks.
Today, insurers have an array of computer programs that guide the flow of trillions of dollars to and from customers around the world. These programs include sophisticated “catastrophe models” that use weather data and other factors to predict an insurance company’s losses in a disaster. “Scoring models” use credit histories and secret algorithms to estimate which customers are more likely to file claims. Colossus and similar programs help companies manage claims. Like a TurboTax program for medical injuries, adjusters plug in information about a person’s loss — from a damaged spine to a fractured finger. Colossus then cranks out a range of payments to cover the costs. Insurance industry critics and even many insiders call these programs “black boxes” because their formulas, data sets and operational policies are cloaked in secrecy.
Romano has thick black hair and wears thin glasses. His brown eyes widen when he wants to make a point. He had gone into the business to help people, but he knew that his work on Colossus would do the opposite.
Birth of Colossus
Among computer programmers, the name Colossus has a rich history. In World War II, British code-breakers called their hulking new programmable machine Colossus and used it to decipher German teleprinter messages. In 1970, filmmakers released “Colossus: The Forbin Project,” a science fiction movie about an army supercomputer that tries to take over the world. (At one point, the computer tells its human creator, “You will come to regard me not only with respect and awe but with love.”)
The insurance industry’s version of Colossus was born in Australia. In the 1980s, a government-chartered insurer ran into financial trouble because of claim costs, which were growing at an annual rate of 14 percent. The insurer set its sights on its adjusters.
It’s the adjuster’s job to evaluate people’s losses and come up with ways to settle their claims. This often meant assessing what people did in their careers and how an injury might affect their future income and overall enjoyment of life. Longtime adjusters talk about the challenge of sizing up people when they’re suffering, and the knowledge adjusters need, from medicine to car repairs, to calculate a fair settlement.
The inherent complexity in putting numbers on injuries also meant that adjusters often came up with different amounts for similar types of claims. In Australia, payments varied by more than 80 percent. So to reduce these disparities and lower overall costs, the Australian insurer worked with a software company on a novel idea: embed the experience and knowledge of their best adjusters in a computer program.
The programmers studied how top adjusters made decisions and then created software to mimic their work. This program became known as Colossus and required answers to as many as 700 questions, ranging from the severity of injuries to how people experienced the loss of enjoyment in life. Injuries were broken down into 600 different codes. The program analyzed legal settlements and jury verdicts, combined this information with data entered by the adjusters, and generated what were supposed to be fair settlements.
In the United States, the insurance industry was experiencing its own period of self-analysis. It began in 1989, when Hurricane Hugo flattened parts of South Carolina. The storm caused $4.2 billion in damage to insured property — at the time the most expensive loss in history. The second wake-up call came in 1992 when Hurricane Andrew generated $15.5 billion in claim payments, $10 billion more than actuaries had predicted. Andrew bankrupted 11 insurance companies and prompted dozens of others to flee the Florida market altogether.
Amid this sticker shock, industry leaders asked why they had so badly underestimated their potential losses. They found answers in newly created “catastrophe models,” computer programs that predicted potential damages in a hurricane or other disaster. These models warned that future hurricanes would be even more costly, and with these new predictions in hand, insurers soon justified massive rate increases in home insurance premiums, especially in South Carolina and other coastal states.
While insurance premiums are the insurance industry’s main source of income, payments for claims are its biggest costs, the equivalent of rubber for a tire manufacturer. Claims also are at the heart of why people buy insurance. Insurance is based on the idea of sharing risk, a grand communal exercise that involves collecting $4.6 trillion every year from people across the world and then shifting some of this to a smaller pool who suffered losses. Insurance keeps communities destroyed by disasters on life support until their economies recover; it helps keep people out of bankruptcy after car wrecks and house fires. And it was largely for these noble purposes that Mark Romano decided to make insurance his life’s work.
An adjuster’s story
Romano grew up in Tampa, Fla., and by his account had a relatively uneventful childhood. He loved catching and dissecting animals for biology classes and thought someday he might go into medicine. He played trombone in the high school band. His mother was a school librarian. His father was regional director for the Florida Department of Agriculture and Consumer Services, and Romano remembers his dad coming home angry about how consumers had been bilked in one way or another. After high school, Romano enrolled in Florida State University, where he gravitated toward the school of insurance and risk management. “I was interested in the basic concept of risk, that you could transfer it from one person to an entity or spread it among many people,” he said. “And you were helping people, and I grew up with two parents who in one way or another helped people.”
Romano’s first job was as an adjuster with American States Insurance, and his first day at work came after a drenching Florida rainstorm. His boss told him to grab a map and clipboard and take measurements of damaged homes. “It was overwhelming, but it was cool,” he said. “I absolutely loved being on the road. Everything was face to face, and it would be very common to meet people in their homes, sit in their kitchen and talk about their injuries.”
Romano handled auto insurance claims and worker’s compensation cases, learned about medicine, the law and how to establish rapport with people in distress. “You did it all, and it was an incredible education in how the world works.” Not all of this education was positive. A year into his career, he took over a new territory, and when he introduced himself to auto repair shop owners, “One guy said, ‘Hey, do you want the same deal as the other guy?’ ” Romano wasn’t sure what to do. “I went to my father and said, ‘These people are offering me things.’ And he said, ‘Don’t you dare ever do anything like that.’ That’s how naive I was at that point.”
Then, after Hurricane Andrew in 1992, Hanover Insurance started closing offices in Florida. It also was a pivot point for Romano. He was mid-way into his career and eager to advance. The place to do this was Chicago, a mecca of property and casualty insurance. He would take a circuitous route to get there, though. He left Hanover and took a job at CNA insurance division in upstate New York, where he learned about a program called Colossus.
By then, the Australian creators of Colossus had sold the program to Computer Sciences Corp., now named CSC, which licensed it to Allstate and many other insurance companies.
CSC’s marketing materials have long touted Colossus as a way to help insurers “establish consistent recommended settlement ranges,” Edward Charlton, a CSC vice president, said in a statement to The Post and Courier. “Without a clearly defined process or framework in place provided by a software tool such as Colossus, claim adjusters may skip important steps or forget to ask pertinent questions of consumers,” he said.
In Romano’s mind, it made good business sense for companies to automate claim payments, though he feared something could be lost without a more personal touch. And based on his years working as an adjuster, the payouts Colossus spit out for CNA seemed fair. He excelled in his job and eventually was transferred to CNA’s bright red headquarters on Chicago’s Wabash Avenue. As he walked into the building, he looked at the skyline. All around were skyscrapers adorned with the names Prudential, Blue Cross, Kemper and Hancock, huddled like giants overlooking Lake Michigan’s southern arc.
The profit center
Allstate was created a year after the stock market crash of 1929, when Robert Wood, president of Sears Roebuck & Co., boarded a commuter train to downtown Chicago. On his ride in, a friend suggested he start an auto insurance company and sell insurance by mail. Wood eventually formed a company called Allstate Insurance Co., naming it after a tire sold in the Sears catalog. In 1950, the daughter of a sales manager came down with hepatitis. When the sales manager returned home, his wife reported, “The hospital said not to worry. We’re in good hands with the doctor.”
Thus, the iconic slogan was born: “You’re in Good Hands with Allstate,” along with the logo of a pair of hands cradling a car. (The car was later removed.) By 2000, the “Good Hands” phrase was the most recognized advertising slogan in America, according to a Northwestern University study. Allstate became one of the industry’s largest insurers, and grew even more in 1999 with the $1.2 billion acquisition of CNA’s personal insurance division.
Romano heard rumors about the deal months before it was made public. A senior vice president approached him and said, “Mark, I hear you know something about Colossus.” The executive told him Allstate was looking for someone to implement their version of Colossus on CNA’s customers.
Allstate renamed the CNA division Encompass, and Romano soon met with Allstate executives who, he said, “began indoctrinating me in their Colossus philosophy.”
Romano discovered that if he used Colossus the way Allstate did, he could save its new Encompass division millions of dollars by “turning the knobs” of the software — paying people less in claims than they would have otherwise gotten.
In South Carolina, for instance, CNA had divided the state into two territories — the “Liberal” area around Charleston and the “Conservative” region elsewhere. Allstate renamed the territories “Charleston” and “Palmetto.” By using Allstate’s Colossus tuning methods instead of CNA’s, Romano could reduce payments in the Palmetto region by 18 percent. Savings were even greater in the Charleston area — a 57 percent reduction. That meant the Allstate version of Colossus would turn a $10,000 claim in Charleston into a $4,300 payment.
“It became my responsibility and goal to save $33 million over three years for Encompass, which I did.” (In a statement to The Post and Courier, Allstate did not dispute Romano’s account but said government regulators have examined its tuning methods and found no violations of state statutes.)
Romano was so successful that Allstate transferred him from the Encompass office downtown to Allstate’s headquarters. Now, instead of downtown Chicago, his commute took him to suburban Northbrook and a 250-acre office park surrounded by fields, security fences and guard gates. “They sent me to the mothership.”
About the same time in 2000, Rob Dietz was working as an adjuster for Farmers Insurance Group in the Seattle area. Like Romano, he felt a sense of purpose helping people put their lives back together. A former logger and rock blaster, Dietz became an adjuster, he recalled, because “it was easier to lift a pen than a chain saw, and because it served the public.” Unlike Romano, he was almost immediately appalled by Colossus.
Farmers was just beginning to implement Colossus. As part of that effort, the company asked Dietz and other experienced adjusters to examine a sample of claims and come up with fair offers to pay people for their losses. These offers would be fed into Colossus to create a benchmark of payouts tailored to that area of the Northwest. But after the group finished, a facilitator said the ranges would then be reduced by 20 percent to create even lower benchmarks.
Dietz was stunned. To him, it meant that the program was being rigged to make payments 20 percent lower than they should have been. “That’s not how I learned the tenets of good faith and fair dealing.”
Worse, after this session, he said he and his colleagues were under constant pressure to stick with Colossus-generated payments even when the adjusters thought people deserved more. He felt Colossus was turning his profession into keyboard slaves, and for a “person with logger’s fingers,” this didn’t bode well for his career prospects. He was also taken aback by the secrecy around Colossus. “I still have the old memo that says we were not to disclose the fact that we were using Colossus.”
Dietz eventually quit Farmers to work with trial lawyers, and in 2002, a Washington State attorneys group asked him and another adjuster to give a talk about Colossus. “That’s when Farmers sued me.”
Farmers asked a judge to stop the seminar, arguing that Dietz and the other adjuster would reveal confidential information. The judge declined, and Farmers eventually dropped the suit. Lawyers from all over the nation flew in for the talk. Aaron DeShaw, an attorney investigating Colossus, remembers how he and the other attorneys gave Dietz and the other adjuster a standing ovation before they opened their mouths. “The atmosphere was electric.”
Good hands, boxing gloves
This was the beginning of what would become a decade-long legal assault on Colossus and other claim-handling programs, one that would somehow bypass Romano, despite his extensive work at Allstate with the program.
One of the most aggressive pushes came from David Berardinelli, a trial lawyer in Santa Fe, N.M., known for his love of vintage Porsches and a book he wrote about his battle with Allstate, “From Good Hands to Boxing Gloves.”
He learned about Colossus while representing a husband and wife hit by an uninsured drunk driver. Allstate refused to pay their medical bills, and curious about Allstate’s hardball legal tactics, Berardinelli sought internal presentation slides and notes about how the company handled claims. In one legal fight after another, Allstate refused to give them up, saying in a court document, it was engaging in “respectful civil disobedience.” At one point, Florida insurance regulators joined the fray, threatening to prevent Allstate from writing new policies unless the company handed them over.
Allstate eventually capitulated, and the materials provided a window into a company in flux. The most incendiary documents stretched back to the early 1990s. At the time, insurers were railing about what they considered a wave of frivolous lawsuits from lawyers who used aggressive advertising campaigns to lure clients. In 1992, Allstate hired McKinsey & Company, a consultant for the nation’s leading insurance conglomerates. One goal, according to a slide, was to “radically alter our whole approach to the business of claims.”
One of the McKinsey presentation slides described how the company could become more efficient if it targeted people who didn’t have lawyers. In its “Good Hands” approach, Allstate would pay those unrepresented people within 180 days, which McKinsey said would take care of 90 percent of the claims. The 10 percent who hired lawyers or didn’t accept claim offers would get the “Boxing Gloves” treatment. In these cases, Allstate would expect to tie up payments for three to five years.
Over time, Allstate employees testified that they were trained to build rapport with customers and discourage them from hiring lawyers. Berardinelli and a growing cadre of lawyers alleged that the “good hands” strategy actually involved delaying and denying claims for several months and then making lowball offers as people felt more financial pressure. They argued that Colossus and other claim-handling programs were important parts of this profit-making plan, with some testimony showing that Allstate could reduce bodily injury payouts by $264 million a year if it used Colossus. “This immediate impact would, of course, come at the immediate expense of Allstate’s policyholders,” Berardinelli wrote in his book.
In a 2008 press statement, Allstate said the materials were part of “a complex body of work that as a whole demonstrates a careful, fact-based analysis to better enable the company to more promptly investigate and more consistently and effectively evaluate claims.” Allstate told The Post and Courier that the software “provides merely a recommendation, and is only one factor in the adjuster’s overall evaluation of the claim.” Charlton, the executive with Colossus’ maker, CSC, said that his company leaves the tuning process to insurers.
Meanwhile, other industry officials have long discounted the importance of the McKinsey documents. Robert P. Hartwig, president of the Insurance Information Institute, said the notion that the documents “forever directed the entire homeowner and auto insurance process” was “bizarre.”
Rather, he said, such programs reflect an understandable use of technology. “There are millions of claims every year and a lot of commonality between them,” he said, adding that said Colossus and Xactimate, a Colossus-like program that handles home insurance claims, “harness the computer to process large amounts of data quickly and inexpensively, and that allows insurers to provide coverage that’s very affordable.” Insurers wage a “technological arms race against each other on a daily basis,” he said, and companies with the best technology have an edge. “This is a competitive industry, and it’s not in the insurer’s interest to treat a customer poorly.”
But Berardinelli and others alleged in class-action lawsuits that insurers were doing exactly that — failing to pay customers what they were due. More documents and testimony emerged, including manuals that described how tuning Colossus was “both an art and a science” that was done “based on the desired projected savings.” One slide from CSC said, “What does Colossus Really do” and begins with a list: “Lowers indemnity payouts ... lowers loss ratios ... improves surplus/profitability.” Other documents urged employees to avoid using the word “savings” to describe the benefits of Colossus and “use a more vague term such as ‘consistency.’ ”
One of the most prominent lawsuits involved a woman from Arkansas named Georgia Hensley. Hensley was driving on a road near Texarkana on New Year’s Eve 2000, when she was struck by an underinsured driver. She broke facial bones and injured her spine. She filed a claim with her insurer, Encompass, which offered $1,000. Hensley’s lawsuit alleged Colossus and other claim-handling programs were cost-containment tools that enhance insurance company profits at the expense of customers.
Hensley’s claim had been handled by one of Romano’s underlings, and Romano was one of the first at Allstate to learn about the lawsuit.
Crisis of conscience
It landed in his email inbox on Feb. 17, 2005. Romano read the lawsuit, a class-action case that named hundreds of insurance companies that used Colossus and other claims-handling programs. He sent it upstairs to the attorneys. By then, he was beginning to feel the weight of his work.
His responsibilities had grown. His tuning directly affected how thousands of claims employees across the country did their jobs, and through them, how much tens of thousands of policyholders were paid for their losses. He was part of a small group of insurance professionals nationwide that met regularly to discuss Colossus-related issues.
These meetings often happened in warm places, including Myrtle Beach. Romano was glad to go to these particular meetings because it meant he could visit his daughter, a biology major at the College of Charleston. They grabbed sandwiches at Groucho’s on King Street and took walks to the Market, where he stocked up on Lillie’s of Charleston Low Country Loco hot sauce, grits and other Southern specialties tough to find in Chicago.
He didn’t talk about insurance, though. The issues he was wrestling with were complex, and he was more interested in how his daughter was doing. He also kept much of his worries from his wife. In 2003, she was diagnosed with multiple sclerosis, and he wanted her life as stress-free as possible. “I didn’t share my feelings about Colossus with anyone, but if I had talked about it, I would have said, ‘I’m doing some stuff that I’m not too thrilled to be doing.’ ”
In his mind, Colossus was as malleable as clay. You could mold its programs to reduce claims values across-the-board, which he described as “turning the knobs.” You could decline to enter data on high jury verdicts or unusually high injury settlements, which tricked the program into thinking an injury’s typical value was lower than it really was. You could train adjusters to code injuries in a way that didn’t account for their true severity, which also reduced payments.
In late 2007 and early 2008, even as the Hensley and similar lawsuits began to produce out-of-court settlements worth tens of millions of dollars, Romano worked on new ways to “recalibrate” and tune Colossus, projects that he said would generally “lower settlement values” and increase profits.
His migraines grew more severe. Doctors prescribed tranquilizers, ordered physical therapy sessions. Nothing helped. He couldn’t sleep. The dizzy spells became more jarring until the doctors told him to turn over his car keys. He temporarily left work and went on disability. Through this haze, he began to see other things more clearly: People were being hurt by Colossus, and it was tearing him apart. He couldn’t turn the knobs anymore.
On his last day at Allstate, he was told to hand in his laptop and badge. On the long drive home, he had no bouts of vertigo, only relief bordering on exhilaration. “It was the first step in regaining my self-respect.” He had a new quest: to help consumers better understand how the insurance industry can fail to live up to the promise of paying people in their times of need. He thought he would be part of a larger chorus, especially now that state regulators had turned their attention to Colossus.
In 2009, led by New York and Illinois, state insurance regulators began the first multi-state examination of how an insurance company uses a software tool to handle claims. Working with the National Association of Insurance Commissioners, the regulators hired a private company to sift through a million pages of claims data and other Colossus-related materials. Investigators later said they spent 8,500 hours reviewing the materials and interviewing more than 40 current and former Allstate employees.
The regulators announced their findings a year later: Overall, they found no “institutional issues involving underpayment of claims” but that Allstate failed to tune the software in a consistent way across the nation. “Colossus was a black box. We looked into the black box and saw some problems,” Steve Nachman, New York’s deputy superintendent for fraud and consumer services, told reporters at the time. “It’s all about how you utilize it.”
Among other things, the regulators ordered Allstate to tell consumers when they had used Colossus to calculate a claim payment. Allstate also was fined $10 million. More than 40 states signed on to the deal, including South Carolina, which received $235,166. (The money went to the state’s general fund.) In a news release, Allstate said the findings showed their use of Colossus “provides significant benefits to the public in increased objectivity and efficiency.”
In a statement to The Post and Courier, Allstate said the investigation in fact justified “the continued use of the tuning criteria which have now been used by Allstate for more than 15 years.” Colossus critics weren’t impressed with the fine or the findings. “Ten million dollars is no big deal,” said DeShaw, the trial lawyer in Washington. “They make that in no time.” (In 2011, Allstate had $32 billion in revenue and a profit of $788 million.)
“A part of this story is the failure of state insurance regulators to police insurance companies’ conduct,” added Jay Feinman, a law professor at Rutgers University and author of “Delay, Deny, Defend,” a book that says insurers try to avoid paying claims.
Robert Hunter, a director with the Consumer Federation of America, was blunter: “It was weak.” If the investigation was so thorough, he asked aloud, why had the regulators failed to talk with Allstate’s official Colossus expert, Mark Romano?
Romano asks himself the same question. The investigation was hardly a secret in Allstate’s hallways, he recalled. He said he even knew where the examiners worked — two miles away near an executive airport. At one point, he contacted an examiner, who told him it was too late to use his information; they had all but wrapped up their work. Romano eventually called Hunter at the Consumer Federation of America.
Hunter remembers the call. “One of the first things he said was that he wanted to help consumers, which is something I liked.” Hunter had already assembled a large body of information about Colossus but was happy to learn about Romano. “Suddenly we had a guy from inside who knew how it worked.”
Romano joined the group and co-wrote a paper last summer with Hunter: “Low Ball: An Insider’s Look at How Some Insurers Can Manipulate Computerized Systems to Broadly Underpay Injury Claims.” It generated numerous stories in insurance trade journals and websites, along with scattered newspaper reports, but Romano acknowledged that “Low Ball” was designed to raise interest among regulators, not the general public, and he’s not sure it made much headway.
These black boxes have a significant impact on what people in South Carolina receive for their claims, but state insurance regulators have no plans to study Colossus or other claim handling programs. They say they leave such analyses to states where insurance companies are based. Overall, said Robert Hartwig of the Insurance Information Institute, “these issues are dead and buried, and regulators don’t pay much attention to it. The fact of the matter, they’re satisfied with the methodologies and constantly review the models.” Twenty percent of the top 30 U.S. insurers, including Allstate, use Colossus today.
Romano isn’t so sure the issue is dead. Insurance is too important to people. He’s seen how it helped make people’s lives a little easier in their time of need. He was proud to call himself an adjuster but knows he lost his way, as has the industry he once so respected. Today, Romano spends his time working on ways to inform consumers about the complexities of insurance, help people the best he can. That’s what he always wanted to do; it’s what insurance is supposed to do. His migraines have all but vanished.
Insurance companies have another controversial black box program that affects what South Carolinians pay for auto and homeowner insurance. Going by “customer rating index” and similar names, these computer models use credit information and other data to estimate whether you are more likely to file a claim. Insurers then use these scores to decide whether to hike or lower your premiums — or deny you coverage altogether.Insurance companies guard these formulas aggressively, so consumers and even regulators have little idea whether they’re being applied fairly.The Post and Courier, for instance, recently asked the state Department of Insurance for “scoring manuals,” citing the Freedom of Information Act. The insurance department then notified State Farm, Nationwide and Allstate about the request and asked for their comments.Insurers demanded that the material not be released, according to emails obtained by the newspaper. “This information is proprietary to State Farm and contains commercially valuable trade secret information that State Farm has collected and created and to which State Farm strictly controls access on a need to know basis,” a State Farm official wrote in one email. “It is understood that absent court order the Department will not release the information produced.”The state Department of Insurance denied the newspaper’s request, even though other states have released these manuals to consumer groups, including the publishers of Consumer Reports. (The newspaper is appealing the department’s determination that the information is confidential.)The result of this secrecy is that consumers have no way of knowing how their credit scores affect their insurance rates. What’s clear, however, is that the issue continues to generate controversy.Insurers cite studies that show people with poor credit histories are more likely to file claims. But many consumer advocates say these scores discriminate against some minority consumers and poor people who otherwise might be good insurance risks.Consumers Union railed against the use of these scores in an extensive study in 2006, saying, “While insurers are preoccupied with gaining a competitive advantage over one another, consumers are getting caught in the crossfire.” Their report found that people could be penalized if they simply opened up several credit card accounts in a year, or made more than two loan inquiries.
This post is an excerpt from The Post and Courier by Tony Bartelme.
email@example.com / 214-293-0944
Office: 469-906-2600 Ext. 101/ Fax: 469-906-2601
9500 Ray White Dr. Ste. 200, Fort Worth, TX 76244