When I posed this question on GovLoop's LinkedIn, Mark Forman gave the following response:
Allison, most government IT systems are modified data bases. Your question is excellent. When you look across government IT projects, you see that the last two decades are characterized by agencies trying to automate knowledge processes using a data bases and strict business rules (basically approve or disapprove; accept or reject a transaction; etc.). There are huge problems for executives in every agency trying to get insights on a problem or issue...
The best example relating to your question is financial information. it takes a month to get anything close to useful on budget status for the President or Director of OMB...versus real-time or near real-time for most commercial organizations. The issue is that there are systems on systems of "proprietary" financial data (meaning only the segment of the agency with data really knows what is happening). As the data get aggregated, they are often mannually re-keyed into different systems, and the real government-wide financial software is not SAP or Oracle, but Excel. There are literally tens of thousands of people across government who reconcile exceptions (aka suspense accounts), and they do their best but cannot reconcile all the financial data (in the $10s of Billions now according to the Treasury Dept). The best data then gets converted from the proprietary account structures to official standard "budgetary" account data, which goes to OMB. OMB compares that to the Treasury Dept data, finds differences and works with the agencies to reconcile them. Finally, OMB gets a good data set to load into its system (MAX), and create a report for the President or other Whitehouse execs. It takes a month...can you beleive it...and it's been going on for years and cannot be fixed as long as agencies maintain thousands of financial-related systems.