Home > Employment > Heritage Foundation’s Misleading Chart

Heritage Foundation’s Misleading Chart

In my ongoing (and largely empty) search to find empirical evidence for the role “policy uncertainty” in slowing hiring, a reader challenged me by sending me this link to a Heritage Foundation report. It was fairly shocking at first how clearly it seemed to show that the passage of Obamacare coincided with the economic recovery “stalling.”

James Sherk writes,

Within two months of Obamacare’s passing, the recovery stalled. Figure 1 shows net private-sector job creation from January 2009 onward. The red line shows the trend in job creation before and after April 2010. Private-sector job creation improved by an average of 67,600 jobs per month before April 2010. That month, private-sector employers added 229,000 net jobs.

As Sherk admits, this is only a correlation not causation but, he writes, “the fact does lend strong weight to the voices of businesses who say that the law is preventing hiring.” It seemed to be at least suggestive evidence of the “policy uncertainty” theory. Yet, my skepticism led me to think about this graph a bit further.

Notice that before around January of 2010 the line graph dips below 0. Although the “trend in job creation” is upward, what that graph actually shows is the economy is losing private sector jobs just at a slower rate from earlier months. Former Bush advisor, Keith Hennessy, had previously criticized Austan Goolsbee for a similar tactic with arrows. (I discuss that video here) The Heritage Foundation even approvingly links that video so they understand what they’re doing.

If you instead look at this bar graph of private sector job creation you can see more clearly what’s going on.

I’ve added in an arrow for April – the month that supposedly shows the beginning of the lower trend for job creation due to Obamacare. But now we see that after Obamacare passed we were gaining jobs while before it we were losing jobs. I’m certainly not claiming causation. I’m merely pointing out the obvious fraud that James Sherk and The Heritage Foundation are trying to get away with here.

[update July 22]: A few other bloggers have noticed them same thing I did about Heritage’s deceitful report. Their pieces are worth reading too. Funny enough, I actually checked all these blogs when I was first sent the Heritage link. Glad they joined the party to expose the propaganda.

Modeled Behavior   Kevin Drum   Matt Yglesias

[update September 13]: I’m happy to see that this post has been cited frequently in numerous online forums. Unfortunately, I’ve noticed that some confusion continues around the change in “trend” in private sector employment before and after the passage of Obamacare. The trend cited by Heritage is purposefully misleading; we were not gaining an average 67,600 before April and then 6,400 afterwards. Repeat: Net change in jobs is not the same thing as job growth. I looked up the private sector job numbers for the dates in the graphs (Jan ’09 – Jun’11) and calculated the average monthly change in private sector employment before and after Obamacare, which is what everyone seems to think the Heritage graph shows.

Before Obamacare the monthly change in private sector employment was an average of about -326,000 jobs a month. That’s NEGATIVE 326 thousand.

After Obamacare the monthly change was an average of +114,000 jobs a month. That’s POSITIVE 114 thousand.

How’s that for a changing trend?

  1. Jack
    July 21, 2011 at 11:50 am

    I’m reminded of how many times I’ve heard criticism of the “lame-stream” media. The truth of the matter is that neo-cons have had to create their own facts in order to insulate themselves from the truth. Reports like this one show that we are not dealing with a think tank. but a propaganda tank.

  2. July 21, 2011 at 1:06 pm

    I am curious as to which sectors were affected. I wonder if it is medical services. If so, we might as well say they are jobs in the public sector since that is the road the medical industry is heading. I know a lot of insurance stock shot up with the introduction of Obamacare.

    I don’t claim to know a lot about this issue, but good job with the bar graph. I think if anything these exercises show how graphs and numbers can be manipulated. I prefer to operate by principles, as do a lot of people, I think. Obamacare boils down to an entitlement benefit, people will decide one way or the other based on that and then find numbers that will support their point of view. So yeah, good job with the debunking.

    • July 21, 2011 at 2:45 pm

      The healthcare industry has been increasing hiring more than most (if not all) sectors. Everyone should always be skeptical of being mislead by stats and graphs. But I worry that people will take that to mean “don’t trust any stats and graphs.” Statistical nihilism is as big a danger as statistical credulity. The best way to combat being fooled is to learn about statistics and economics and to always read competing views.

  3. July 21, 2011 at 4:59 pm

    That’s true. I, of course, am guilty of blowing off graphs if they don’t fit into my paradigm and quick to peddle ones that I agree with, but for the most part I try to take them all in to paint a bigger picture. I think that they are very instructive though, and should lead you to dig deeper. They are also some factors that can be graphed that are so clear cut that there can be little debate as to what is going on.

    But I digress. As far as the issue at hand, that bump s pretty definitive. I would like to see someone come up with some other reason that it did that.

    Just watched the graph comparison video and your deconstruction. Good stuff. I think the most convincing argument Hennessy brings up is that there were a lot of factors. The one that struck me the most was time. In his revised graph as well as your bar graph it looks like overall the economy has been following a long trend that looks suspiciously like the normal business cycles. It is something that will happen over and over. Whether Keynesian economics is making the valleys shallower and the peaks higher is an almost century old debate. Meanwhile, no matter what, politicians will cast blame for the bad times and claim responsibility for the good times. It’s the same with these trends, generally unprovable.

    As such, I think spikes, like the one we see on your graph are generally more informative, but even there it seems obvious (to me at least) that it had little effect on the general trend.

    As an aside, job growth has begun to grow less fast. I don’t discount the fact that it is growing, but I do find it disconcerting that it has leveled off so quickly and looks like it is headed toward losing jobs again. I hope that on a longer time line ends up looking like an outlier.

  4. July 21, 2011 at 5:13 pm

    It’s very worrying. And politicians in both parties seemed to have stopped trying. The reality is that employment situation has hardly improved at all – we’re treading water basically: http://research.stlouisfed.org/fredgraph.png?g=1dA

    I think we should do something about the long-term debt as much as the next guy, but not at the expense of ignoring our current economic problems. The biggest disappointment in all of this is the Fed, which doesn’t have the same hurdles to action that our political institutions have. A better monetary policy could do a lot to speed our recovery.

    • July 21, 2011 at 5:21 pm

      I totally agree. Fiscal policy is a sledge hammer that takes at least six months to wind up and strike down whereas the effects of monetary policy can be felt the same day.

      That being said, I am all for a deal that will cut debt, and cut spending any day of the week. Call me old fashioned, but I start feeling sick to my stomach when I think about how much money gets thrown away on interest payments. I expect the same fiscal responsibility from my gov’t as I practice in my home.

      • July 21, 2011 at 5:26 pm

        Argh, I must diverge. The government and a home are too different to analogize between. I don’t want to get too off topic from my original blog post so I’ll leave it there. Maybe I’ll cover that in another post.

  5. July 21, 2011 at 5:41 pm

    I am just saying that there are general principles that should be used to practice fiscal responsibility. If you need to leverage some debt for awhile to improve conditions, that’s fine, but the who idea of gov’t spending (to me at least) is that you spend when you have to, but tighten the belt during times of plenty to replenish the stores. The problem i have with most gov’t spending programs is that they are so often entitlement programs that don’t go away and so when the next downturn happens, we are caught with our pants down.

    So I looked that graph (http://research.stlouisfed.org/fredgraph.png?g=1dA) and if you don’t mind me throwing a correlation out there, I find it interesting that the drop-off coincided with unemployment extensions. Funny how when less people are working the economy tends to dive. But I don’t really want to get into a chicken or the egg debate. Obviously there are bigger players then some benefits, but I can’t see how that would help.

    My actual theory is that there are too many people play WoW and living in their parents’ house.

  6. Jack
    July 21, 2011 at 6:18 pm

    What is clear is that wages have been stagnant since the mid 1970’s and that income inequality is increasing. I attribute this to the weakening of the new deal policies that have been occurring since the 1970’s. We need to return to policies that promote the middle class and not some top down coalition that we been pursuing for the last 30 years.

  7. Ococker
    August 15, 2011 at 5:09 pm

    How can one legitimately compare net change in jobs with private sector employment?
    Are these not actually two distinctly different cohorts?
    And what is the latter? Since April 10 4 of the 16 months have seen “private employment” exceed 200,000. Really?

    • August 15, 2011 at 5:35 pm

      The 2nd graph is also showing the change in private sector jobs. It shows “monthly change in thousands.” I can compare them because they are showing the same thing – one is a line graph, the other is a bar graph.

  8. Ococker
    August 15, 2011 at 9:46 pm

    Respectfully, I think not. The bar graph plots change, not the net change depicted in the line graph.

    • August 15, 2011 at 11:05 pm

      Feel free to look at the numbers in each graph by month: they’re the same numbers. Each graph is illustrating the change in private sector jobs each month.

      You’re welcome to play around with the raw numbers and graphing tools at the St. Louis Fed if you don’t believe me. Plug in the dates, alter the “units,” and then switch back and forth between bar and line graph and you’ll see the same trends (with more up-to-date numbers) we see in the two graphs in my post.

      Thanks for reading.

  9. Ococker
    August 16, 2011 at 10:40 am

    You are, of course, correct- the data are the same. However, the line graph’s rate of change add a flavor that the other doesn’t.
    But enough. Thanks for tolerating my amateurish thoughts.

    • August 16, 2011 at 9:00 pm

      No worries, just glad I could clear everything up. Hope you keep reading whether you disagree or not.

  10. Dean Moriarty
    June 1, 2012 at 12:08 am

    “I’m merely pointing out the obvious fraud that James Sherk and The Heritage Foundation are trying to get away with here.”

    Gawd it is so nice to see that at least SOMEONE is keeping an eye on those freaking CHARLATANS. The fact that the “research” that comes out of those think tanks is used without hesitation as support for arguments heard on talk radio, Fox News, and on Sunday Talk Shows makes me wanna pull my hair out. Great work Dan.

  11. JP
    October 24, 2012 at 7:04 pm

    By changing the graph from one that shows the acceleration of job growth per month to one that shows the total amount of job growth per month, you obscure the relevant point, but you don’t negate it. You can still see, from your chart, that the rate of job growth, the rate of change in the height of your bars, is increasing sharply month-to-month, then, at the point of Obama’s healthcare law, it suddenly drops, then flat-lines. This is the very same trajectory that is shown in the Heritage Foundation’s graph. Fit a line that follows the peaks of the bars until April 2010 and you have a line that’s rising at about a 45% angle, a very healthy and robust recovery. Fit a line that follows the peaks of the bars from April 2010 on and you have a nearly flat-line rate of increase, the stagnant employment growth that we have today.

    Where you are confused, or intellectually dishonest (I don’t know you so I won’t speculate which), is where you put the focus on the number of jobs created per month rather than on the acceleration of the number of jobs being created per month. Obviously there are more jobs being created per month after the healthcare law passes than at the depths of the recession. But the critical point is the speed with which new jobs are being added, that it is accelerating rapidly from the depths of the recession until the passage of Obamacare, and then acceleration virtually stops at that point.

    You simply cannot compare the job losses at the depths of the recession to the modest job gains after Obamacare. An accurate measure of the rate of recovery, one that can apply to both negative and positive job growth, is the ACCELERATION of job growth. And this acceleration is clearly greater before Obamacare than after, and significantly so. Therefore, your point that there are many more jobs added per month after the law’s passage than before is completely missing the point. If we were adding 10 new jobs per month after the passage of Obamacare, by your logic, we would still have been doing better than we were before Obamacare passed, when we were rising rapidly from steep losses per month, by virtue of the fact that we were losing jobs each month before, and gaining jobs, however few, after. That alone should illustrate the flaw in your reasoning. Had the pre-Obamacare trend continued, the total number of jobs added per month should have gone MUCH higher in a relatively short time, and unemployment would have reached an acceptable level long ago.

    I also find it unfortunate that you, and others who visit this site (and reference it), use this “rationalization” to conclude that conservative “think tanks” are merely propaganda. It, unfortunately, calls into question your own objectivity.

  12. October 24, 2012 at 9:33 pm

    Thanks for your response. I promise, no effort at intellectual dishonesty here. It seems you’re mistaken in your analysis however. We can see why the Heritage graph is so misleading; it’s just done so for you. Take a look at the “rate of job growth” in the last graph I post. That’s the same job figures in a different format. As you can see, pre-Obamacare there wasn’t an “acceleration” of job growth there was a deceleration of job loses. Hence the wide “U” shape. The early trajectory of the Heritage chart only goes upward because they move the line up through negative numbers. Remember: every point on that graph below zero is represents jobs lost not gained.

  13. JP
    October 24, 2012 at 9:39 pm

    I’m a physicist. Deceleration and acceleration are the same thing. One is the negative of the other. I can understand your confusion.

    You’re focused completely on the wrong thing. You should be looking at the rate of change in job growth, not the number of new jobs added or lost. The rate of change early on is very large, even though it’s a deceleration of job loses. Do you understand this? To stop losing jobs at such a drastic rate, you must “accelerate” in the opposite direction, or “decelerate”, until you reach zero losses, then you begin to accelerate into positive growth. In every graph you show, there is an obvious change in the rate of acceleration at that inflection point, where the law is passed. It is beyond dispute. The only question is, what has caused it? That can be debated.

  14. JP
    October 24, 2012 at 9:40 pm

    Oh, and thanks for responding and engaging in the debate!

    • October 24, 2012 at 10:02 pm

      I get that the pure rate of change is different pre and post obamacare. I’m not disputing that, but it’s totally misleading for the Heritage Foundation to 1) suggest a causal relationship 2) to suggest that we were gaining more jobs per month pre-Obamacare.

      • October 24, 2012 at 10:31 pm

        Let me just try to clear this up further. As a physicist, you must realize that Sherk of Heritage just took the second derivative, which isn’t the most useful calculus function for measuring meaningful economic trends. Although we, of course, are all just matter in motion responding to the laws of physics, economics doesn’t use the same models that physics does. For good reason!

        If we were wondering why the speed of a ball changed its trend we might have a useful question, but let’s examine why expecting the same rate for “net change in private sector jobs” leads to absurd results. To quote economist Karl Smith: “Suppose that the pre-ACA trend had continued unabated. Then on top of the surprisingly strong job growth of April 2010 we would have kept accelerating at a rate of 67K jobs per month. At that rate we would currently be adding jobs at roughly 1.2M per month every month.”

        No economic model would predict a trend to continue that added 1.2 MILLION jobs every month. And to assume that a healthcare bill passing accounts for the difference between 1.2M jobs/month and 6,400 jobs/month is pretty implausible.

        But I really appreciate being challenged on this and other topics. I hope you continue to read and comment.

  15. JP
    October 25, 2012 at 12:18 am

    “I’m not disputing that, but it’s totally misleading for the Heritage Foundation to 1) suggest a causal relationship 2) to suggest that we were gaining more jobs per month pre-Obamacare.”

    I don’t agree that they did. It is certainly debatable whether there is a causal relationship, and that debate would be a better use of our time. And I don’t see where they suggested that we were gaining more jobs per month pre-Obamacare. They state that we had a greater acceleration (or deceleration of losses) pre-Obamacare.

    You recognize that there was a change in the economic data at that point. Fine. So the question is, what caused that? That’s worthy of debate. But the point of this post is to deny that there was even a change, and that’s false.

    As to your other point, most recoveries “snap back”, with the speed of recovery proportional to the depth of the recession. Such recoveries take on a “v” shape. That is the shape this recovery took early on. Rarely, there are recoveries that take on an “L” shape, and these are sometimes called a “lost decade”. An example of this is the Japanese recession of the 90s. The shape post-Obamacare is that of an “L”-shaped recession. The data now is worse than what you see in this graph, in fact, with only a 300 jobs/month rate of growth. So while I think you’re right, in general, that the rate of growth of jobs/month can’t be sustained forever, I think it’s also true that it should have been sustained for a longer period, giving us a more robust and healthy “v”-shaped recovery. Something happened at that point, whatever you want to attribute it to, that caused the expansion of the economy to slow drastically, and prematurely, before the economy had recovered to the point that it normally would have.

    So that’s really all I’m trying to say. That there is clearly a change there, that the rate of expansion of the economy flat-lined after a brief period of more robust recovery. And I’ve looked at graph after graph, all showing various metrics about this recession, and there is always this inflection point where the trajectory of the recovery changes in a negative way. And that’s all the Heritage Foundation graph purports to show, and your post seems to deny that. I’ll leave the debate about what caused it to others.

    • October 25, 2012 at 2:48 am

      I’m still not sure why you think a trend in how many fewer jobs we’re losing would continue at precisely the same acceleration as jobs created. Look at that last chart from the Fed again. We were losing jobs. We had less total jobs each month.

      Say we have 100 total jobs. In the 1st month we lose 30, in the 2nd month we lose 20. Now we have 50 total jobs. So, we had a slow down of the rate businesses were laying people off. Say we’ve hit bottom and firms start hiring again. What about the prior “trend” makes you think they’re going to hire back at the same rate they were laying people off at? Again, the economy isn’t a ball in motion.

      The Heritage chart should be labeled +67,600 jobs/month^2, not doing so and not providing a clear explanation of the difference misleads readers into thinking each month saw a net gain of 67,600 jobs – that didn’t happen.

      Finally, let me just point out that “the speed of recovery proportional to the depth of the recession” is not always true. As many economists such as Carmen Reinhart and Ken Rogoff have pointed out, financial crisis are different than normal recessions. Compared to other financial crisis recessions and recoveries, we aren’t doing so bad:

      Of course, I’m happy to argue along with everyone else that we should have done more and we could have done much better, but there is no evidence that the ACA played any major role in hampering or slowing our recovery. To suggest otherwise is post hoc ergo propter hoc.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: