The President's Council of Advisors on Science and Technology (PCAST) has studied government cybersecurity, and it says the picture isn't pretty.

"The federal government rarely follows accepted best practices. It needs to lead by example and accelerate its efforts to make cyberattacks more difficult by implementing hardening practices for its own systems," reads the 17-page, recently released report from PCAST. The report is a declassified version of a study that PCAST -- which is comprised largely of members of academia and major technology vendors -- submitted in February 2013 to President Obama.

Some of the PCAST report's recommendations include getting government agencies to ditch Windows XP within two years, increase their use of trusted identities, share threat intelligence sharing with the private sector, and design systems that are hardened against attacks from the outset. "Research is needed to foster systems with dynamic, real-time defenses to complement hardening approaches," according to the report.

But how many of PCAST's recommendations promise the best bang for the government's cybersecurity buck? "This report reads as being very disconnected from what the federal government is actually doing or has been doing, and what has worked or hasn't worked," said John Pescatore, director of the SANS Institute, in a recent SANS newsletter. For example, the report only makes a passing reference to the Federal Risk and Authorization Management Program (FedRAMP), which gives government agencies cost-effective, standardized techniques for -- in the program's own language -- "assessment, authorization, and continuous monitoring" of cloud-based services. The goal is to only have to conduct an assessment once, then to let government agencies subscribe as they see fit.

Pescatore added that the PCAST report also omits any mention of the Department of Homeland Security's Continuous Diagnostics and Mitigation program, which DHS is building "to provide adequate, risk-based, and cost-effective cybersecurity" for the executive branch. That would seem to square well with PCAST's call for better "dynamic, real-time defenses" for the government. In the future, furthermore, DHS plans to extend the program to include the defense industry, as well as state and local government agencies.

Do federal agencies really need the PCAST report's recommended trusted identity program? Would the report's call for greater information sharing between government agencies and the private sector lead to better federal cybersecurity? "It was very obvious that the report was written mostly by folks from universities, along with Craig Mundie of Microsoft and Eric Schmidt of Google: many recommendations for more government funding for research, government updating of operating systems and browsers, and more use of cloud services," said Pescatore.

Where is the business case?
Given the PCAST members' vested interests in many of their recommendations, it's notable that -- despite many of them hailing from the private sector -- they've failed to offer a business case for any of their suggestions.

For example, what's the return on investment to be gained from PCAST's recommendation to advance the trusted identities program? For comparison's sake, the Department of Energy hasn't been able to get the Congressional appropriations it needs to replace an outdated -- and recently hacked -- Adobe ColdFusion system that stores personal information on every employee and contractor. But fixing that vulnerability would immediately plug an existing security hole, protect identity information, create a safe foundation for building future agency applications, and demonstrate an immediate return on investment for an agency that's now paying for credit monitoring services after losing information on more than 100,000 current and former employees, contractors, and their dependents.

Or take the PCAST recommendation to replace Windows XP, despite the use of the operating system remaining widespread not just in the government, but also the private sector. Microsoft, of course, has been warning people to upgrade before it ceases related support come April 2014, and there are strong security reasons to do so.

But should government agencies pay to upgrade to Windows 8, which will necessitate purchasing new hardware? Instead, why not instead install virtualized -- and locked down -- Windows XP clients, thus retaining existing hardware and avoiding many refresh costs? That combination might not appeal to PCAST members, but it would provide better information security at an affordable cost, and avoid the expense of silver-bullet thinking.

Another useful take on PCAST's recommendations comes via Gal Shpantzer, an information security and risk management advisor who offered -- also in a recent SANS newsletter -- the following finding: "It is important to influence designers of future computers and software so that security controls can be installed before the fact and as an integral part of the system. It is also important to ascertain what can be done with equipment presently installed or owned by the government."

That recommendation isn't from the PCAST report, but rather the Defense Science Board Task Force on Computer Security. As the group's name might suggest -- note the lack of the government-speak word "cybersecurity" in the title -- the report isn't recent. In fact, it dates from 1970. Perhaps the more the government's information security problems apparently change, the more they really stay the same.

Mathew Schwartz is a freelance writer, editor, and photographer, as well the InformationWeek information security reporter.