So, lots of people have been discussing strange, strange claim by Heritage that Ryan's budget would drop the unemployment to 2.8% by 2021. (A claim which they later revised to 4.2%, by the way.) But it does not seem that anyone has fully understood what happened here. So let me explain.
First, you need to know a few things.
* There is a concept in macroeconomics called, somewhat confusingly, "full employment unemployment rate", or, more accurately, "non-accelerating inflation rate of unemployment" (NAIRU). You can read all about it in Wikipedia. In the US today it is thought to be around 5%. CBO assumes that, by the end of this decade, it will stand at 5.2%.
* In macroeconomic forecast models such as those used by CBO and Heritage, it is an external parameter that you put in manually.
* In the long run (like, say, 10 years), the unemployment rate in these models tends to converge to NAIRU.
* There's no reason to think that NAIRU under Ryan would be substantially different from NAIRU under Obama.
Now let's get back to Heritage. They were asked to model the US economy under the assumption that Ryan's budget is adopted. This means accounting for changes in government spending, taxation, etc. Among other things, they had to account for changes in workforce.
Heritage is a conservative organization. One of the cornerstone principles in modern conservative economics is that taxes are bad, because they disincentivize people. Capital gains taxes disincentivize investment, and personal income taxes disincentivize labor (therefore pushing people out of the workforce and lowering the labor participation rate.) Personally, I think that it is a bit suspect (when was the last time you heard of anybody quitting a job because they couldn't tolerate the brutal 28% marginal income tax?) But for conservatives, that's a crucial concept that can't be ignored.
So their analysts plugged the numbers, ran the simulation ... and then there was a little problem. Turns out that Ryan's plan does not reduce the effective income tax on the middle class, on the contrary, it increases it. (Just the repeal of the mortgage interest deduction amounts to a 5% to 10% effective tax hike on many homeowners.) And that, in their own model, translates into a shrinking work force.
If you look at their original spreadsheet, you can easily calculate that, in their model, adoption of Paul Ryan's budget shrinks the work force by about 2.2 million people as of 2012. Smaller work force at the same level of unemployment means fewer people on payrolls, and that means smaller GDP.
At that point they could've gone forward with the announcement. "Paul Ryan's budget decreases the ratio of publicly held debt to GDP in 2021 from 107% to 65%, destroys 1.5 million jobs, and lowers the GDP by 0.5% compared to baseline." Do you see that happening? Me neither. Their corporate overlords wouldn't be pleased. And, while they could do something about GDP, it's quite hard to mask the decrease in employment.
And then one of the analysts probably had a bright idea. They set NAIRU to 2.8% and redid calculations. Artificially increasing employment by 2.5% magically turned net job loss into net job gain. The report was sanitized of all negative information. (Do you find it curious that the report, which explicitly mentions labor participation rates in their methodology, fails to include them in the spreadsheet?) They would've removed the unemployment row too, but it was too important to claim a rapid drop in unemployment, and they didn't want to mess with the data directly, for some reason.
Of course, once their results were thoroughly ridiculed in the blogosphere, they had to go back and retract. The way that retraction was done is telling, too.
* They ONLY release the updated unemployment rate, but not payrolls or GDP. They do assure us that other results of the simulation "do not change significantly." Riiight.
* They STILL don't hike NAIRU back to 5.2% where it should be, but only to 4.2%. Why? I suspect that I know why. 4.2% is the highest level where they can put it without falsifying their original claim of an increase in private employment through 2021...