Numbers Game

The IT budget proposal contains modest increases, but it might understate spending.

Travel out West and chances are you'll see cheat grass, though ideally you shouldn't. An invasive species that earned its name by choking out native species such as sagebrush, it also has a propensity to burn easily.

Look layers deep inside President Bush's $64.3 billion request for federal information technology programs in fiscal 2007 and you'll spot a fractional request-for $100,000-by the Bureau of Land Management for a weed database. The idea is to map and then anticipate invasive species before they infest new areas, BLM says.

But funding for the weed database is likely to wither. And BLM isn't alone. On the whole, Bush's request represents an increase of less than 3 percent over the amount Congress appropriated for IT in fiscal 2006-a rate below the Office of Management and Budget's inflation estimate. It's less than the $66.5 billion, in inflation-adjusted dollars, the White House sought for fiscal 2006.

Rather than attempt to replicate the ambitious fiscal 2006 request-which in inflation-adjusted dollars amounted to an increase of 5.5 percent over the previous year's request-administration officials are resigned to a lower base line, says Karen Evans, OMB administrator for e-government and IT. "You can't just continuously put money down in a bad investment," she says.

Congress usually appropriates less IT funding than the administration requests anyway, although information technology doesn't always have its own line item for appropriators to specifically target. Many in federal IT circles like to keep it that way; obscurity is a useful shield against appropriators' swords, they argue.

Still, despite a history of curtailing administration expectations, Congress generally allows the IT budget to grow slightly from year to year even when inflation is factored in. The small increases and cuts happen because many lawmakers consider IT an overhead burden rather than an enabler of agency missions, says Paul Brubaker, who had a large role in shaping the 1996 Clinger-Cohen Act, which overhauled federal information technology management. Until the government can demonstrate clear results emanating from IT, the budget picture is likely to remain the same, says Brubaker, now executive vice president for SI International in Reston, Va.

Tracking outcomes directly attributable to IT investments is difficult. Technology-created upsurges in productivity often aren't measurable until years after the technology's initial introduction, according to a bevy of economists. Lawmakers "get as frustrated as anybody over the perceived lack of progress," Brubaker says

Plus, IT is "not usually the kind of thing that gets constituents hot and bothered," says Stan Collender, a veteran budget observer and columnist for National Journal (a sister publication of Government Executive). Lawmakers would rather spend money on "a dam, a post office, a tunnel-something they can see or feel," he adds. As a result, IT contractors and federal agencies should view the president's request as extremely tentative, he says.

But the proposed IT budget could be a poor reflection of reality for other reasons, too. "The real cost of IT is a hell of a lot more," says Robert Charette, former chair of the Software Engineering Institute's risk advisory board and president of ITABHI Corp., a Spotsylvania, Va.-based risk-management consultancy. The numbers in the president's request aren't a true reflection of the actual effort being poured into the management and support of IT, he says.

The Defense Department, for example, does not report expenditures for hardware and software embedded in weapons. Extracting those costs from weapons system budgets would be virtually impossible, says Defense spokesman Lt. Col. Brian Maka. Even at civilian agencies, IT is so entrenched in daily operations that real costs likely outstrip budget projections. In the private sector, every dollar invested in IT generates as much as $5 more in training and business process reorganization costs, says economist Robert J. Gordon.

There is a less tangible reason for the mismatch between reality and budgets, too-a pervasive culture of penalizing honest estimates of project costs, Charette says. Getting funding for a project often requires overpromising what can be done with limited resources, he adds. And once a project is successfully funded through unrealistic estimates, a disincentive ripple effect discourages full disclosure. The problem worsens during tight budget times, such as now, Charette says.

In unveiling the budget, OMB says it improved management of IT programs governmentwide. The number of major IT projects on the agency's watch list for underperformance in metrics, earned value management and cybersecurity fell to 263, from 342 last year. But large projects are getting larger. The value of major IT projects is projected to increase to $37 billion in fiscal 2007, but there are fewer of them. As a result, the average value per major project will have increased by 44 percent in inflation adjusted terms since fiscal 2005.

Reducing project size to improve management was a main tenet of the watershed Clinger-Cohen Act. Nonetheless, there exists "this love affair with these megasystem procurements," Brubaker says. People assume that tackling complex problems requires complex solutions, but complexity adds risk, he adds. Agencies tend to stuff budget requests with large projects because they're averse to breaking up acquisitions into smaller, modular components for fear funding will dry up, Charette says.

But the trend toward larger projects could be a good thing, says Mark Forman, Evans' predecessor at OMB, now a KPMG LLP partner. A more focused effort to manage fewer projects is better than many smaller projects in which the systems integration isn't well thought out, he adds.

OMB officials say recent improvements in their oversight make IT project failure less likely than it once was. Annual agency business cases and earned value management ensure that projects have stated goals and milestones. "And yet, the success rate of projects, especially large scale projects, hasn't changed very much," Charette says.

A recent Government Accountability Office review (GAO-06-250) of 29 business cases submitted by five agencies each proposing to spend at least $1 billion on IT during fiscal 2006 found problems with the process. One agency submitted a nearly decade-old risk management plan as part of its business case. The required analysis of alternatives in almost all the business cases lacked documentation, or contained conflicting evidence, GAO said.

Nonetheless, the idea that greater accountability leads to better funding has been established, Forman says. Brubaker agrees. "At least [OMB] is asking some good questions and forcing some thought, which before really didn't happen."

But as long as project managers are penalized for delivering bad news, OMB's oversight will be ineffective, Charette says. OMB officials have said it's OK to be the bearer of bad tidings. "People out there in the field hear that, but they don't believe it," and experience confirms their belief, he adds. Travel through federal agencies and you'll likely see a fair amount of game-playing with IT budget numbers, even though you ideally shouldn't, Charette says.

The fiscal 2007 budget request includes $5 million for the OMB-controlled e-government fund. In addition, the administration wants permanent authority to use up to $40 million for e-government from funds generated by General Services Administration fees. President Bush made the same request for fiscal 2006, but Congress appropriated only $3 million and rejected the idea of using any GSA fee money for e-government.