Why do governments make such a mess of IT?

by John Naughton

The Observer

This is a tale of two cities — Washington and London — and of the governments that rule from them. What links the pair is the puzzling failure of said governments to manage two vital IT projects. In both cases, the projects are critically important for the political credibility of their respective administrations. And yet they are both in trouble for reasons that most engineering and computer science undergraduates could have spotted.

So here’s the puzzle: How is it that governments stuffed with able and conscientious civil servants screw up so spectacularly whenever IT is involved?

Let us start with Obamacare, the U.S. president’s landmark reform of his country’s dysfunctional healthcare system. If Barack Obama is going to have a place in history other than as the first black American to reach the Oval Office, then his Affordable Care Act will be the achievement that does it. The legislation had an agonisingly tortuous passage through a rancorous and partisan Congress, but eventually it made it onto the statute book.

Implementation of Obamacare depends on a website that enables U.S. citizens to shop online for health insurance. Introducing it on Oct. 1, the President said “Just visit healthcare.gov, and there you can compare insurance plans, side by side, the same way you’d shop for a plane ticket on Kayak or a TV on Amazon. You enter some basic information, you’ll be presented with a list of quality, affordable plans that are available in your area, with clear descriptions of what each plan covers, and what it will cost. You’ll find more choices, more competition, and in many cases, lower prices — most uninsured Americans will find that they can get covered for $100 or less. And you don’t have to take my word for it. Go on the website, healthcare.gov, check it out for yourself.”

You’ve probably heard what happened: thousands of people tried to do just that and found the site didn’t work. For the first few weeks only a fraction of those who wanted to sign up for an affordable healthcare plan succeeded. It was, in a word, a fiasco.

In the rolling inquests that followed, institutional factors were the first things to be singled out. It was claimed, for example, that the companies contracted to do the site were more expert at negotiating the 1,800-page federal procurement manual than at designing web services. And it also turned out that Obama’s advisers were so paranoid about Republican attacks that they refused to allow the beta testing essential to debug any high-traffic site.

But actually the real reason why the site was unworkable was architectural: It required applicants to open an account and let the site verify their identity, residence, and income before they could browse for insurance. That meant the site would have to interface — in real-time — with databases maintained by the Internal Revenue Service and other agencies. Since those databases all run on their own timescales, and some are pretty slow, that meant that healthcare.gov could only be as fast as its slowest link. It would have been much smarter to allow anyone to browse without needing to verify their identity, and only require authentication when they were signing up to a healthcare plan.

The strange thing about this is that you wouldn’t need to have been a geek to spot the problem with healthcare.gov. You just had to think architecturally about it. Yet apparently nobody in the administration did. The same applies to the post-9/11 decision to link all the previously separate U.S. security databases into one giant file to which at least 250,000 people had access, one of whom happened to be Bradley (now Chelsea) Manning.

Meanwhile, on the other side of the pond, Britain has its own IT mess — the computer system underpinning universal credit, the government’s flagship reform of the benefits system. Like Obamacare, it’s a politically critical project but parliament’s Public Accounts Committee has taken it apart. The committee’s report lambasts the project’s “extensive delay,” not to mention “the waste of a yet to be determined amount of public money.” The PAC thinks that much of the £425 million (¥71.3 billion) already spent “will now have to be written off.” The management of the programme has been “alarmingly weak.” And so on.

The message here is the same as with the Obamacare site. This is not about computer science or technology, but about common sense and management. The good news is that we could therefore do something about these fiascos. The puzzle is why we are taking so long to get that message.

  • fuscator

    The type of failures described here, and the pervading mediocrity of projects that do not outright fail, stem from the fact that, as a rule, government employees do not “pay” for failure, nor are they especially rewarded for performance above expectations.