That process now takes, on average, at least 61 days. That’s the same amount of time it took in 1977, according to a federal audit from that time. Many state retirement systems, which also handle large loads of employees, do it much faster. Florida takes 47 days. The California teachers’ retirement system takes 23. Texas takes two.
Those three process their files digitally, not on paper. Since the 1980s, the U.S. government has been trying — and failing — to do the same thing here.
The first time, work began in 1987. Years passed. About $25 million was spent, according to the Government Accountability Office. But within the government, officials started to worry that it wasn’t working.
“The reports [from the contractor] just asserted that they had written X lines of code. . . . For an executive, that’s just invisible; you don’t know what it means,” said Curtis Smith, who oversaw retirement processing from 1989 to 1994. He was a longtime federal employee with a PhD in English literature, supervising a massive technology project.
“I had no idea [if] they were making progress from month to month. And I just sort of took it on faith that they could make it work,” Smith said. “And they never did.”
In 1996, two years after Smith left the government, officials finally pulled the plug on that project. Then, in 1997, the government tried again.
First it tried revamping the system in-house. Then it scrapped that plan and hired contractors. After years of work, the system the contractors built was supposed to be ready by early 2008.
But by 2007, there were concrete warnings that it again wasn’t going to work.
“Every time we would do what I would call a stress test, we would come up with abysmal numbers — like an 18 percent success rate,” said Robert Danbeck, who was overseeing the project. The root of the problem, he said, was that the system had trouble synthesizing information from so many sources and calculations based on so many laws. “We would go back and look at what caused it, and it was always just so many pieces, trying to tie things together.”
Danbeck quit. In early 2008, the system went live.
Then it broke and was eventually scrapped, after more than $106 million had been spent. In the mine, the files continued to move on paper.
Contained in all those failures, experts say, is a very brief history of the federal government’s recent troubles with information technology.
A recent study by the Standish Group, a firm in Boston that researches failures, found that only 5 percent of large federal IT projects in the last decade fully succeeded.
Of the rest, 41 percent were failures, canceled before they were turned on. The reasons often echoed the problems in the mine: Federal officials either tried to buy a technology they didn’t fully understand because they lacked the technical skill, or they didn’t test what they were getting until it was too late.
I love reading the reports on how long it will take to synchronize U.S. military accounting systems. Some estimates say 100 years which is another way of saying that we have given up and aren’t even trying any more.