If you, gentle reader, have not had the opportunity to be outraged and personally insulted by MIT Economist Jonathan Gruber’s statements on “the stupidity of the American voter” and “a lack of transparency” making it possible to get Obamacare passed, you may have been out of the country, suffering a bad case of the flu, or relying on the mainstream media for news for the past month.
Gruber, a one-time darling of the left, created the now infamous GMSIM, or “Gruber Microeconomic Simulation Model,” a computer model upon which Obamacare, and before that, Romneycare, was built. Quite a few states have also paid Gruber hundreds of thousands of dollars for his healthcare modeling expertise.
The GMSIM model uses economic theory combined with actual experience to predict the results of any proposed health insurance law.
Obamacare is a law. Why does it need a Model?
A simulation model is simply a set of mathematical equations that takes data measured in the real world and manipulates it to predict what will happen in the future. A good model can predict with accuracy, while a poor model’s predictions usually won’t come true.
For example, a good weather model uses data about actual temperature, humidity, wind speed and direction, etc., over a large geographic area, and runs a computer “simulation” to predict what temperature you can expect tomorrow, and the chances that it will rain.
A good model can predict with accuracy, while a poor model’s predictions usually won’t come true.
According to GMSIM documentation on the MIT website, Gruber and his team of grad students used data on 80,000 individuals and families, collected in the census, along with data from insurance companies, the non-profit National Bureau of Economic Research (NBER) and the Bureau of Labor Statistics (BLS).
This part of the model is pretty straightforward, as long as you trust the data coming from these organizations—which is not a question to be ignored, but let’s assume it’s OK for now.
Added to this real-world data, the economists input “varying information of policy parameters.” With this, they can predict the effect of different policy changes by converting changes in the law to estimated price changes. Or, in other words, it can predict how much insurance will cost for different groups of people, based on how the law is written.
Is Gruber’s Model Accurate?
Anyone can make a model, but the question remains, is the model accurate? An accurate model will predict the future accurately. An accurate weather model will predict the temperature a day or a week from now, and sure enough, the temperature will register fairly close to what was predicted. If not, the model sucks.
If the model does NOT accurately predict the future, modelers will often try to fix the situation using “fudge factors.” As defined by Google a fudge factor is “a figure included in a calculation to account for error or unanticipated circumstances, or to ensure a desired result.” (italics mine)
Computer modelers of all kinds throw fudge factors into their models
Computer modelers of all kinds throw fudge factors into their models in order to get their simulations to match what has happened in the real world. Fudge factors are one way the scientists running the simulations can make them come up with the answers that they want.
Another way to adjust the model to get the “right” answers is to change the starting point, ending point, or duration of the simulation run. For example, in modeling the costs of Obamacare, the clever lawmakers looked at a 10-year window of time, but deliberately delayed implementing some of the provisions for two or more years, so that the cost would be appear lower.
“Torturing” the Model
Gruber, who teaches Public Finance and Public Policy at MIT, knows how to “adjust” his model in order to change the outcome quite drastically.
In one of his infamous videos, he admits that “this bill was written in a tortured way to make sure CBO did not score the [individual] mandate as taxes. If CBO scored the mandate as taxes, the bill dies.” He understood that if it became clear that the mandate was a new tax imposed on them, the public would revolt.
This was unacceptable, politically, so Gruber and his “grubs” went to work to get the model to come up with a completely different answer. Of course, none of this was done in a closet. Congress actually “loaned” Gruber to the CBO to “help” them score the Obamacare law properly, as Max Baucus admitted on the Senate floor.
They discussed how to surreptitiously kill the health care tax benefit
Gruber is on record meeting with the head of the CBO in the Oval Office with President Obama, during which they discussed how to surreptitiously kill the health care tax benefit in such a way that no one would notice or complain.
After several months of messing with the model in order to get the right answers, in March 2010, CBO declared that Obamacare would become “deficit neutral” in 10 years, meaning that the costs of implementing the law averaged across that time would equal out the savings to the Federal Government. However, the law would only be in operation 6 of those 10 years, making the scorecard a joke.
The CBO’s finding, however tortured, was crucial in getting support for the bill in Congress, and was used as a cudgel to beat in the heads of conservatives and Republicans for months.
CBO, Other Models All Based on Gruber’s
According to a 2012 article in the New York Times Gruber spent decades developing his proprietary GMSIM model. Writer Catherine Rampell says that it was Gruber who convinced the administration that Obamacare could not work without the individual mandate (Don’t forget that Obama had run in 2008 on the position that a mandate was unconstitutional).
Futhermore, Rampell wrote, Gruber “has nearly cornered the market on the technical science behind these sorts of predictions.” She also quotes a different economist who claims that other similar models “all use Mr. Gruber’s work as a benchmark.”
Gruber is the only person you can go to for that kind of thing
“He’s the only person you can go to for that kind of thing,” Rampell quotes Harvard Economist David Cutler as saying, “which is why the White House reached out to him in the first place.”
In fact, the Obama Administration had no choice but to hire Gruber: his model is an inaccessible black box, due to its proprietary nature.
Persona non Gruber
But, perhaps most distressing for the way forward, is that Gruber’s GMSIM model, with its deliberately tortured and deceptive results, now forms the basis to which any future changes, updates or replacements to the Obamacare law will be scored.
And because the CBO doesn’t reveal how it calculates the costs of a particular bill, there is no way of telling exactly to what they will compare any future healthcare bills.
But according to J. Anderson writing in the Weekly Standard, “because the model that the CBO used in scoring Obamacare is the same one it uses today, any alternative to Obamacare that doesn’t include an individual mandate — which is to say, any conservative alternative — would be scored by the CBO as falling well short, in terms of coverage numbers.”
Gruber himself has quickly become persona non grata in the Democratic Party
Even though Gruber himself has quickly become persona non grata in the Democratic Party, his GMSIM model lives on, spreading deceit and mayhem wherever it is found.
But how does this cautionary tale of a computer model run amok relate to Climate Science? We’ll discuss that in the next article.
NEXT UP: Climate Science in the Age of Gruber, Part II: Models, Models Everywhere