Wednesday's Rocky Mountain News featured a fascinating look at the latest in a string of information technology failures in Colorado state government: "Doesn't compute: 'It's like you were having a baby, and it turned out ugly'." The article describes a failed effort to develop a new motor vehicle licensing system, CSTARS, for the Colorado Department of Revenue (DOR). The project is a perfect example of how too much of a good thing can end up wasting massive amounts of the taxpayer's money, how privatization is useless if it doesn't yield meaningful competition.

Like many ambitious projects, CSTARS began in 2001 with optimistic planning, and a contractor that was willing to promise everything the agency asked for. A beautiful baby was expected by all. Two years later, the project became bogged down in problems as the contractor discovered how difficult the job really was. Sensing high risk and impending trouble, and unwilling to admit to the customer that it was in over its head, the contractor began cutting corners and shedding functionality. It took another four years of management denial before a new administration came along, saw that the project had become a slow-motion trainwreck, and began to pull the plug. How did this happen?
DOR had assembled a wish list of 700 requirements, and put them out to bid. Of the seven firms that responded, six had developed similar systems before and felt that not all of the 700 requirements could be satisfied at the same time for the available budget. The seventh firm, Avanade, had not developed a similar system but promised it could satisfy all 700 requirements. The state accepted the bid even though Avanade clearly had not provided sufficient evidence that it understood the job. DOR had not forced the contractor to prove its ability to manage the work. In fact, as it turns out, DOR was also incapable of managing the work.
Bill Owens' former Revenue Director was quoted in the article as claiming CSTARS isn't the same kind of computer project failure that Colorado government had experienced four other times under Owens with four different contractors, most famously the $223 million Colorado Benefits Management System. He's wrong. It is exactly the same kind of failure. It's also the same kind of failure that caused an improperly-secured bridge girder to fall onto I-70, killing a family in a minivan almost exactly three years ago. You might think these are unrelated technology problems, but they aren't: in every case the technology involved was relatively simple and straight-forward.
What went wrong in all of these projects is the political mistake of confusing privatization with competition. It's a common mistake that causes a lot of bad policy decisions. Economists have long understood the value of competition as a source of actionable information. When a project is put out to bid, the prices that come back tell an agency what the best available price might be, and who is offering it. This can then lead to the action of writing a contract favorable to both the agency and the contractor, based on the winning bid. Usually the private sector supplies most or all of the bidders; hence the term privatization.
But the bid preparation process is not only a source of essential information, it also consumes and transforms information given to it. Like any computer program, it's garbage in/garbage out. A list of 700 features is not enough information to give a reliable cost estimate on a large information system. Without a preliminary functional design analysis, DOR did not supply enough information to expect meaningful bids to come back. Effectively, it asked the bidders to "figure it out," which is not much different from asking them to take a wild guess. Having been on the bidder's side of that equation many times over 27 years, I recognize the setup. Without a preliminary attempt to design the system, DOR didn't know that it might be impossible to implement as conceived. Without its own systems analysis and internal cost estimate, it didn't have a framework for evaluating the competence of the bids.
The reason DOR didn't do its homework, is clear in the article. It failed because it didn't have the necessary manpower and expertise on its own staff, to make sure it was getting usable information in the bidding process. Further compounding the error once the job was awarded, DOR didn't have the expertise to understand the software project workflow, nor the importance of the functional design process. It didn't exercise control over changes in requirements, so it lost control of the project. It gave too much responsibility and power to the contractor, and didn't protect the public interest. An understaffed agency is easily fooled.
This then is exactly the same problem that occurred on I-70 with the failed bridge girder. The Colorado Department of Transportation had given most construction supervision responsibility to the contractor, and was not staffed to supervise the work adequately. A family died as a result.
Privatization is a competition not only among bidders, but between the agency and the winning contractor. The agency and contractor don't have the same goals in mind. The DOR wants to get a computer system that meets the needs of its staff and the county clerks, so they can best serve the public in the issuance of license plates. The contractor wants to make a profit, win more projects with other clients so it can grow, keep its staff busy, and stop the client from complaining about being behind schedule. Those are competing goals. If the agency is too weak to compete, too understaffed to gather the information it needs, it won't know enough about contractor capabilities, project status, or achievement of project milestones and objectives. It won't be able to enforce the contract. A contract that isn't enforced, isn't worth the paper it's written upon. If the DOR doesn't have the staffing to effectively represent the public interest, it will end up wasting the taxpayer's money.
Privatization can be a good thing, if it provides real competition. Competition is a good thing, if it provides useful information and the agency acts on it, to its advantage. But stupidly cutting back on staff and throwing money at the private sector, just because of blind faith that "the private sector does everything better," is a huge mistake. While private firms can do a great many things better than public agencies, one thing they will never do better is to look out for the public interest.