I’ve been asked many times about the discrepancy in reported and published numbers presented by agencies such as the Congressional Budget Office (CBO). People want to know, for example, how one day federal health reform is reported to save the US money and the next day cost billions.
Unfortunately for me, there is no simple explanation as to how these varying estimates are produced… and the lengthy, statistical clarification quickly kills any dinner party I’m invited to.
The jaded, academic in me wants to claim that in the end, researchers can make numbers say anything they want to; that the outcome is more closely related to the questions you ask or the lens through which you view the results. However, this is ultimately untrue. The real reason numbers vary in reports is due to several factors including, but not limited to, using different thresholds, double counting, and rounding.
A great illustration of this is the Federal Poverty Level (FPL) thresholds used by the CBO when analyzing and projecting Medicaid estimates. The current health reform bill has a 5% income “disregard” built in, meaning that the first 5% of someone’s income is ignored when determining how far above or below the FPL a person falls. Therefore, Democrats use estimates based on 133% FPL and Republicans use estimates based on 138% FPL. This may seem like a small amount of discrepancy until you recall that we are talking about health reform effecting millions of people and billions of dollars in state and federal funding.
Maybe the real numbers we should be paying attention to are who is paying and how much for the calculations being reported. This will help determine which base numbers are being used and which direction the rounding will favor.