Widgetized Section

Go to Admin » Appearance » Widgets » and move Gabfire Widget: Social into that MastheadOverlay zone

Why Does the Federal Government Have So Many Old Computer Systems?  

The views expressed are those of the author and do not necessarily reflect the views of ASPA as an organization.

By John Pearson
March 12. 2018

The Government Accountability Office (GAO) and others have written about the need to modernize the federal government’s legacy IT (information technology) systems. Here’s a link to a list of 10 very old federal computer systems. Some are over 50 years old.

President Trump signed the Modernizing Government Technology (MGT) Act on December 12, 2017. TechRepublic.com reported on December 13, 2017, that, “The MGT Act creates a $500 million fund over the course of two years to be used for modernizing legacy IT systems.”

The US government currently spends about $80 billion a year on IT, and some estimate that 80% of that is spent solely on maintaining old systems.”

The MGT’s $500 million over two years sounds like a drop in the bucket to address the problem of the government’s legacy systems. Because of funding limitations, I believe retention of these old systems is a perfectly rational response to the situation federal agencies find themselves in.

Near the end of my career, I was the business liaison for several legacy systems. (In my agency, we always referred to the program side of the agency as “the business side” to differentiate from the IT side.) None of these legacy systems were simply databases used for analysis. All were rules based systems that to some extent replaced employees or enhanced what employees were doing. Even systems going back to the 1970s were like that. Such systems are critical to agency operations and I believe are a major reason the federal government has been able to maintain level employment of about two million employees over many decades.

The older legacy systems are not still around today because of an organization’s “culture” as some have suggested. They are not around because of incompetence or neglect. The main reason for retention of these old systems is the high cost of replacing them.

Development of new IT systems can be a very costly proposition. For example, healthcare.gov, the website supporting Obamacare, reportedly cost nearly $840 million per secretary Burwell’s testimony on May 8, 2014. I recall one system at my agency had development costs of around $250 million.

Replacing a legacy system can be like developing a new system. In my experience, it was rare that a legacy system could be replaced and modernized by the IT department or contractor alone. Usually, a replacement system had to be redesigned from the ground up. Complete requirements for the new system had to be written.

The business knowledge to write the requirements often was hard to come by. Computer systems perform a subset of all business requirements for a given problem. The average business employee doesn’t have the full perspective of what the business requirements are, let alone the ability to define what the computer can efficiently do versus what the human employees can do.

Even if a contractor is used to “elicit” the business requirements, this is a slow and complex process. So-called “agile” development techniques can be used to speed the development process but there still is need for hundreds or thousands of pages of various types of requirements documents – business requirements, functional requirements and technical requirements. Programmers must program the requirements and testers (both from the IT and the business) must intensively test every aspect of the system to be sure there are no defects or the business can live with the defects. The entire process can easily burn millions of dollars for just one IT system, and the federal government has thousands of systems.

The legacy systems, in many cases, are performing pretty well. From a business standpoint, there may not be much value in spending large amounts of money to replace these old systems. I’m not sure it’s true the older systems are more vulnerable to breakdown than the newer systems. IT systems are not like cars.

Every year, an agency must consider its IT budget. What will be spent on maintenance of legacy systems? What will be spent on new systems to address pressing business problems? What will be spent on efforts to replace legacy systems?

The IT department naturally concentrates on spending money to maintain the existing systems to avoid system failure. Most IT dollars go into maintenance. System failure could have a major adverse impact on business operations. The business side naturally wants to address pressing business problems by bringing new systems online or by making modifications to existing systems.

Replacing an old system may provide little or no improvement to business operations. Neither the IT nor the business side has a strong motivation to replace the old systems if they are performing adequately and don’t pose too high a risk of failure.

I heard year after year that we have to replace the older legacy systems. But I saw only slow progress toward that goal because of the problems outlined above.

Putting off replacing the old systems each year is a rational, well-considered decision. It’s the result of agency managers trying to maximize utility for the total IT dollars that are available each year


Author: John Pearson recently retired from a lengthy career in the federal government where he was a program analyst. He has an MPA and a bachelor’s degree in economics. He now writes columns reflecting on his experience in government. His email is [email protected].

1 Star2 Stars3 Stars4 Stars5 Stars (4 votes, average: 4.00 out of 5)
Loading...

2 Responses to Why Does the Federal Government Have So Many Old Computer Systems?  

  1. john pearson Reply

    March 18, 2018 at 5:08 pm

    Yes, I agree with Clyde that sometimes agencies try to replace older computer systems and the replacement project fails. I should have mentioned that in the column. I don’t agree that the project failures are the result of consulting companies wanting to have “butts in seats.” I recall the contractors were under tremendous pressure to produce results. I recall contract employees pulling “all night” efforts to meet deadlines. In the failed projects that I observed, whether the failure was in-house employees or contractors, the project complexity simply overwhelmed the employees assigned to the project. A major problem was the need to define a realistic solution to match the available resources. If you try to accomplish too much (sometimes referred to as scope creep), sometimes nothing useful gets done.

  2. Clyde Poole Reply

    March 12, 2018 at 6:28 pm

    Having worked for and with government agencies for many years, my perspective is slightly different. I believe that most legacy IT systems in government are there because repeated attempts to upgrade/retire/rewrite/replace result in costly failures. The failures are primarily the result of project management being outsourced to consulting companies, where the primary goal of the consulting company it to ensure that any new project takes as long as possible and involves as many consultants as possible. “Butts in Seats” and not successful project completion is the goal. Agency employees are not the problem you will note, but the tendency to try to control agency head count by not hiring permanent employees to manage and build-out these projects.

Leave a Reply

Your email address will not be published. Required fields are marked *