Driving the 480 miles from Columbus, Ohio to Raleigh, N.C., should take you about eight hours. But what if your GPS device insisted you could make the trip in four? Blindly following such an unreliable estimate, even when made by your trusted GPS, would cause you a lot of problems.

Congress too often makes the mistake of blindly following projections of the Congressional Budget Office that later prove to be grossly inaccurate. The CBO’s reputation among the public and the media may be strong, but its track record in providing accurate estimates to Congress leaves much to be desired.

This isn’t a new problem. Scholar Alan Reynolds noted clear back in 2001 that CBO’s flawed projection model often leads to inaccuracies, pointing out several inaccurate CBO projections in the 1990s. He noted their budget projections widely missed the mark even though they were made only 12 months in advance.

Longer-term projections can be problematic as well. CBO’s cost estimate of the 2002 Farm Bill was off by $137 billion. Six years later, they miscalculated the cost of the 2008 Farm Bill by a whopping $309 billion.

Beyond short- and long-term budget projections, CBO has also proven grossly ineffective at forecasting consumer choices. Nowhere is this more clearly seen than the Affordable Care Act.

As the ACA made its way through Congress in 2010, policy decisions were made based on flawed CBO projections, such as the one saying 21 million Americans would enroll in the insurance exchanges by 2016. The actual number was less than half that, around 10 million, and the miscalculation was disastrous for consumers.

CBO also miscalculated other economic impacts of the ACA. For example, they projected Medicaid expansion to cost $4,200 per enrollee. The actual cost turned out to be to $6,366 per enrollee. Likewise, CBO projected a 3.2 percent annual GDP growth under Obamacare from 2010-2016. The actual economic growth rate during that time period was 2.1 percent.

Again, both errors directly impacted budgets and consumers.

The point is, CBO economists are trained to apply Keynesian economic formulas to predict future consumer behavior, but their projections don’t often account for common-sense realities, such as the wet blanket effect that Obamacare had on job creation and economic growth. Time and time again, their projections miss the mark because they don’t consider economic reality.

Fast-forwarding to today’s debate over replacing the ACA, the CBO has once again weighed in with a doomsday projection about how many people would lose coverage if either the House or Senate reform bills pass.

Unfortunately, we believe the CBO has missed the mark again.

First of all, the accuracy of their numbers has been called into question. As Foundation for Government Accountability experts noted in Forbes recently, the CBO projected around 19 million Americans would buy insurance on the individual market in 2018 should the BCRA become law. This would amount to a 7 million person reduction in insured as opposed to what would happen under the current law of the ACA, based on the CBO’s 2016 baseline projects.

However, there’s a significant problem with this analysis: the CBO updated its baseline projections in January 2017, based on the unexpectedly low ACA enrollment figures. Again, the FGA notes that the revised CBO projections for the ACA predicted that 19 million Americans would buy insurance in 2018 under Obamacare.

In other words, the updated CBO baseline for Obamacare showed that the exact same number of people would buy insurance in 2018 under the BCRA as would under Obamacare, a figure which would be one million more than the 18 million currently on the individual market.

More importantly, their projections miss the basic point that people are more likely to purchase an affordable product that meets their needs. The fundamental flaw of Obamacare was that it tried to force a one-size-fits-all plan on consumers that prefer choice. Reforms like removing penalties, lifting mandates, allowing consumers to shop across state lines and increase their purchasing power, will result in more people being insured than the CBO is projecting.

Healthcare reform needs to happen. Hard-working Americans deserve better than Obamacare, which was built on lies and forces too many people into coverage they can’t afford and don’t want.

We also believe it makes sense for Congress to explore the idea of reforming the culture of over-reliance on the CBO. We still need independent, nonpartisan analysis to hold our government accountable, but a thoughtful bipartisan review of this issue may help us find a more accurate alternative to the CBO. Such a review should also consider whether dynamic scoring models will provide a more accurate projection of the economic effects of public policy.

The value of having outside experts review legislation cannot be understated. But continuing to hinge congressional actions on the projections of an agency that has proven to be so consistently wrong does a disservice to not only members trying to represent their constituents, it primarily does a disservice to the public.

The GPS is right quite often. But when it mistakes a boat launch ramp for a road and tells you to “turn left here,” you have to stop and take a look at reality before blindly following the command and turning into a lake.