Don't believe the hype

Anything coming out of a "think tank" that has a very obvious partisan affiliation should be taken with a grain of salt.  It's not a guarantee that the numbers will be twisted, but if something looks wrong... it's probably wrong.  Brad DeLong from Berkeley points out a perfect example of this.  The Heritage Foundation put out a "report" which claimed that jobs mysteriously stopped appearing after the passage of the Affordable Care Act (health care reform).  And the nice graph they put up looks pretty impressive, too.  It shoots pretty far up, and then settles into a flat pattern.  Aha!, Heritage says, this proves that ACA destroyed job growth  Well, not quite.  Take a look at DeLong's link, and you see what the actual job numbers were.  It looks... more or less like a parabola.  Steep on the end, bottoms out, then rises a little bit.  But the rise kind of dies and starts looking more like a constantly sloping line than a parabola.

So what does the Heritage graph measure? Well, it shows the SECOND DERIVATIVE of an employment graph.  In other words, if jobs are being shed, but they're being shed more slowly than they were the month before, then the second derivative is positive.  On the other hand, if jobs are being added, but more slowly than they were before, the second derivative is negative.  ACA passed about 2 months after employment bottomed out.  It spent about two months adding jobs, then kept adding jobs after ACA, but at a constant rather than accelerating pace.  What does this mean? Well, to me, it looks a lot like this.  Fall of 2008 was the panic phase of the financial crisis.  No one was sure if we'd have an economy-- Lehman collapsed in September, and Citi, Merrill Lynch, AIG, Morgan Stanley, and even Goldman Sachs were all in danger of collapsing at some point in those two months.  By the time the graph starts (at the beginning of 2009), all of those institutions had been acquired, bailed out, or turned into bank holding companies by the Fed.  But jobs typically follow output gaps; that is, the economy didn't instantly lose every job in the recession as soon as Lehman collapsed; unemployment didn't go from 5% to 11% overnight.  Instead, it was a gradual process.  Eventually, once the Fed and Treasury had stopped the bleeding, some jobs were regained.  But, given the size of the gap the crisis left, and the debt overhang that workers and businesses now had (from their frenzied borrowing in the first 8 years of the decade), recovery was never going to be quick, vigorous and rapid.

So, what the data says is essentially this: Before ACA passed, unemployment was getting worse, but it was getting worse more slowly than it had been earlier.  After ACA passed, unemployment was getting better, but it was getting better more slowly than it had been getting worse before ACA.  Heritage wants you to believe that this should be taken as a sign that ACA killed job growth.  Looks unbelievably stupid to me.  And should to anyone whose mind isn't made up before looking at the data.

The lesson here is: If something comes out of Heritage, look at the data very closely.  It's not automatically complete BS.  But the chances are about 99% that it's complete BS.