In a recent post, I cited the GAO estimate that the total cost of the Ground-Based Midcourse Defense (GMD) national missile defense system, projected through FY 2017, was now $40.9 billion (in FY 2013 dollars). I thought it would also be interesting to look back over previous GAO reports to see how the cost of the GMD system has increased over the last decade.
The $40.9 billion figure was reported in the GAO’s March 28, 2013 version of its report on Assessments of Selected Weapon Programs. The GAO began releasing this annual report in 2003, although the GMD system was only included starting with the 2004 report. Each report includes a “Latest” cost figure, which is total cost of the GMD system including future costs not yet incurred, projected several years ahead.
The GMD costs appear to include the GBI interceptors and their fire control system, the Sea-Based X-Band radar, and the GMD upgrades to the Cobra Dane and early warning radars, but (probably) not the forward-based TPY-2 X-band radars. It is unclear (to me) to what extent operations and sustainment costs are included (in at least one year they are explicitly excluded). Figure 1 below shows the GMD costs for the ten years the GAO report has been released, starting with the 2004 report (click on the graph to get a larger image).
Figure 1: GAO Total Cumulative Cost for the GMD System, including some future projected costs.
On Figure 1, the x-axis is the date of the data used in the cost estimate. For example, the first point, from the 2004 report, is based on data up to February, 2003 and is thus plotted as x = 2003.17. The y-axis is GAO’s “Latest” cost projection, which is in dollars corresponding to the year of the report and includes costs projected several years ahead. For example, the $22.5 billion point for 2003 is in FY 2004 dollars for costs through FY 2009. The last point is from August, 2012 and is for costs through FY 2017. However, costs associated with the March 15, 2013 announcement of plans to increase the number of deployed GBI interceptors from 30 to 44 are not included in that total.
There are several problems with Figure 1. First, the numbers of years each annual estimate projects ahead varies from year to year. Second, inflation is not taken into account. Figure 2 below attempts to compensate for these two problems by subtracting out the projected funding and by then converting the remaining costs to FY 2013 dollars. Figure 2 is thus a plot of the amount spent on the GMD system through the date shown in constant FY 2013 dollars.
Figure 2: GMD actual costs (no future projections) in FY 2013 dollars.
If we assume the zero point for GMD costs is February 1996, which the GAO takes as the starting date for their cost estimates, then as Figure 3 below shows, the data fits surprisingly well to a linear increase over time.
Figure 3. GMD actual costs fitted to linear plot with a February 1996 starting date.
Nevertheless, it is clear from looking at the ten plotted data points that the rate of GMD spending has slowed up somewhat in the last four or five years relative to the five previous years. This is hardly surprising, since much of the system was deployed by 2010 (for example, the 30th GBI interceptor was deployed in September 2010).
 For the 2005 and 2006 estimates, the GAO gives projections for both FY 2009 and FY 2011. The FY 2011 figures are used here, since these are more consistent the numbers of years projected ahead used in other years.
 The correction for inflation is based on data from table 10.1 of the U.S. Budget for Fiscal Year 2012, Historical Tables (http://www.whitehouse.gov/sites/default/files/omb/budget/fy2012/assets/hist.pd). The chained price index (which the table states is what used for constant dollar research and development outlays) was used.