15 Feb 2014

A Red Herring? Substitution vs. Income Effects on CBO Analysis of ObamaCare

Health Legislation, Shameless Self-Promotion 39 Comments

[UPDATE below.]

My latest at Mises Canada. Key excerpt:

For example, if the government says to workers, “We’ll give you $50,000 in annual support payments, so long as you don’t ever take a job,” then this would obviously be a disaster. I think Yglesias would agree wholeheartedly with me.

Yet suppose instead the government said, “We will randomly pick ten percent of citizens from the phone book, and we will provide each of them with vouchers entitling them to $50,000 worth of housing, food, and car leasing. The citizens are free to earn additional income if they want; they will be given these vouchers regardless of their official wage and salary income.”

This second scheme would also lead to a huge reduction in work effort offered by the 10 percent selected citizens, and it would be through an income effect: They would quite rationally choose to work fewer hours, and enjoy more leisure, because the government was now providing many of the items that they otherwise would be buying with their labor earnings.

One might argue that the first scheme is more disastrous than the second, and in this sense a given expenditure of taxpayer subsidies is less harmful if done in the form of income effects. But even so, the second scheme is still bad (if you agree that the first one is).

I do not claim to know the future, but I am quite confident that Daniel Kuehn in the comments will argue that Yglesias said nothing wrong.

UPDATE: Over at EconLog Scott Sumner has a similar post. He quotes from Jonathan Gruber who makes an important distinction between “voluntary” and involuntary reductions in labor supply emanating from the ACA, but–as Sumner explains–these categories are both voluntary from the individual worker’s perspective. Sumner et al. aren’t using the terms “substitution effect” and “income effect,” but it’s comparable to my discussion about Yglesias and Douthat.

39 Responses to “A Red Herring? Substitution vs. Income Effects on CBO Analysis of ObamaCare”

  1. Silas Barta says:

    Daniel Kuehn: “Well obviously, Yglesias is aware of all that, and if you squint *really* hard, you can see how is writing is not completely inconsistent with having absorbed that insight.”

    • Major_Freedom says:

      “Well, such a simplistic way of putting it would be something he doesn’t altogether disagree with, however there are more caveats than that, which complicate the issue, and while I don’t know for sure, it’s likely that those caveats are what he is leaning towards, and so you’re probably talking at cross purposes, and Yglesias is making a valid point, albeit a more sophisticated one, which I will criticize him for not making more explicit, so all in all, “Red herring” is a bit of an overstretch, but I won’t disagree with your reasons for why you would have been lead to believe that Yglesias is making the point you attribute to him, but in my personal view, the key issue is not that he is making the point you think he is making, which I think you will agree is not exactly the way you describe it, but rather, that when it comes to issues such as these, it is not unreasonable to assume that if you believe he is making a silly point, then chances we don’t have all the facts, because if you go by his previous writings, he’s been much more clear on this, but then again, I myself have made the point numerous times that he could do better explaining things, so in that sense, but not completely in the way you might believe is the case, I might be inclined to agree with the general point you are making, but only if the point is the narrow one that says subsidizing people makes them less inclined to work, but this must be put into a context that we’re not in a liquidity trap, because then demand effects might dominate the income effect, which may be, although he isn’t making it clear, the point Yglesias was making all along, but as I said above, he could have been more clear on that.”

  2. Jonathan Finegold says:

    I only read the excerpt, not the whole article, so I might be saying something that’s already addressed. But, I don’t see why the second scheme is “bad.” It has a cost: a reduction in total labor hours supplied. But, to really be able to say if the policy is bad we have to compare the benefits, and people might think that the benefits are worth the costs.

    • Bob Murphy says:

      Right Jonathan but I said several times they are either both bad or both good. If you spell out why you think the ACA’s reduction of labor supply through substitution effects could “need fixing” (or whatever Yglesias said), then those same reasons would hold true for income effects.

      • Jonathan Finegold says:

        I probably need to think about this harder. I didn’t read Yglesia’s posts (I hardly ever read Yglesia’s posts). But, in the first case in your example, what seems to me the biggest problem is the mandate to never get a job. If people were to voluntarily reduce their supply of labor hours that’s one thing, but we’re talking about a mandate to reduce their supply of labor hours.

      • Jonathan Finegold says:

        Ok, thinking about it a little harder, I think I might see your point (thinking about it in terms of an ugly looking budget constraint).

      • Daniel Kuehn says:

        Can you explain how you are using substitution and income effects here Bob?

        Are people using these words to mean “relative price changes” and “lump sum transfers”, respectively???

        • Ken B says:

          Second this request.

          Just don’t do it with logarithms.

  3. Steven Landsburg says:

    Don’t forget that under the second scheme, the reduced work effort by the 10% is offset by the increased work effort of the 90% who foot the bill for these transfers.

    • Bob Murphy says:


      Right, but the same is true in the first scheme, assuming taxes are financed the same way.

      I think the big problem here (not talking about you, of course) is that people actually aren’t thinking through WHY it’s bad if the government literally pays people not to work.

      • Transformer says:

        “people actually aren’t thinking through WHY it’s bad if the government literally pays people not to work.”

        An alternative framing of “Society decides to create a safety net to make its members feel more secure and this has the unavoidable side-effect of creating a slight distinctive to work” makes the same thing sound slightly less evil.

        • Tel says:

          The idea of the “safety net” is that it only gets offered to people who genuinely can’t work and therefore the marginal effect on employment is zero. That’s the idea, at a theoretical level.

          Whether it really does that is another matter and probably a big argument in itself.

          • andrew' says:

            That’s not what a “safety net” is. But close enough for government work. The current safety net keeps getting closer to the opposite of a literal safety net. Tight rope walkers are working and taking risks.

    • Gamble says:

      Hi Steve,

      What if the 90% lose motivation and do not work any harder?

  4. Transformer says:

    How are you defining “disastrous” ?

    Obviously the 10% will be better off and the other 90% will be worse off/. Austrian are barred from making interpersonal utility comparisons so there is no way of telling if its disastrous or not.

    • Daniel Kuehn says:

      Intelligent human beings are “barred from making interpersonal utility comparisons”.

      • Transformer says:

        Is Peter J Hammond a non-intelligent Human being ?


        • Daniel Kuehn says:

          Note the fourth word in the abstract and notice how significant it is for the rest of the abstract.

          I am guessing I don’t have much to disagree with in the paper as long as he hews pretty closely to that.

        • Daniel Kuehn says:

          You are absolutely entitled to decide that we want to think about welfare as if everyone had the same (or at least similarly scaled – obviously preferences are different) utility function. I personally think this is very reasonable, and it’s more or less how I think about it.

          I cannot say that person X actually values a dollar as much, twice as much, half as much, or k-times as much as person Y. It’s possible neuroscientists may one day be able to make claims like that, but I won’t speculate. That’s very different from what we’re doing in the first paragraph.

          Hammond makes this point very nicely in the last sentence of his abstract: “Instead ICU’s need to be recognized for what they really are — preferences for different kinds of people.”

          • Transformer says:

            Yes, that all makes sense.

            I was actually just trying to annoy Bob with my initial comment – I suspect my views on ICU’s are pretty much the same as yours (and his).

            • Daniel Kuehn says:

              I thought at first you were just trying to yank his chain until you responded, then I wasn’t sure so I spelled it out!

              • Transformer says:

                Second response was just to annoy you 🙂

                But actually your next response was good because I took a look at the paper (that I had just googled) and learned something.from it.

        • Daniel Kuehn says:

          Put simply, not being bullied by utility monsters is not the same as thinking that utility monsters don’t exist.

  5. Daniel Kuehn says:

    I’m curious what makes you think I’d say that about Yglesias? What in my view of labor supply or income or substitution effects do you think is so different from yours that would lead you to think that?

  6. Daniel Kuehn says:

    I don’t think the Sumner post was quite comparable to Yglesias. I don’t agree with Yglesias, but I didn’t agree with Sumner’s post either.

    The best I can figure is that Yglesias is thinking about the lump sum nature of a transfer like that (in theory). But ti’s not theory, the transfer in the ACA IS a function of income, and even if it weren’t I don’t see how the nice features of a lump sum transfer mean there’s nothing to be concerned about with an “income effect”.

    It’s kind of sloppy language to begin with (was there more detail from Yglesias or just this tweet?). Every price change has both an income effect and a substitution effect.

  7. Transformer says:

    More seriously I think the issue may be this:

    If you believe its a good thing to subsidize healthcare that people previously had to pay full price for (or got as a work benefit) then you have to accept that this will result in people working less.and still think its a good thing.

    However if you implement it so the subsidy gets reduced as people earn more then this acts the same a higher marginal tax rate and adds an additional distinctive to work.

    For example say I earn $0 and get health care worth $1000. Someone offers me a job paying $10 an hour. I’m more likely to take it under a scenario where I keep both the $1000 health benefit and the $10 than if the health care benefit tapers by $5 for every $10 I earn.

    It seems that you could be consistently in thinking that health benefits are a good thing (even tough it means lees hours worked) while implementing it with a taper is bad (because it unnecessarily created a marginal disincentive to work). I see no contradiction.

    • Yancey Ward says:

      then you have to accept that this will result in people working less.and still think its a good thing.

      None of this would be an issue if people like Gruber and Krugman had written exactly what you just wrote-namely that ACA would still be a good thing despite the disemployment effects. However, that isn’t what they wrote, is it?

      • Transformer says:


        They make it sound like people working less is a GOOD thing in itself. ACA is a market distortions that will (other things equal) cause people to work(and produce) less than without it. You had better be sure the benefit of this distortion are worth it before you implement it.

        • Transformer says:

          The first effect is an unavoidable effect of subsided healthcare but (its supporters would say) the benefits of universal healthcare outweigh it.

          The second effect (marginal income) is avoidable and therefore bad

  8. Bob Murphy says:

    Daniel, I can’t prove this, but especially in conjunction with Sumner’s post about Gruber, I *think* this is what Yglesias and Douthat are trying to say:

    “If the CBO thinks the ACA will cause workers to reduce their hours so much, primarily because before they were working to get health care, and now they don’t need to do that…then it’s great. But if workers reduce their hours because the subsidies are tied to income and thus there’s a ‘notch’ in which the marginal return to additional labor income is really low, then it’s a bad thing.”

    • Transformer says:

      The first effect is an unavoidable effect of subsided healthcare but (its supporters would say) the benefits of universal healthcare outweigh it.

      The second effect (marginal income) is avoidable and therefore bad

    • Ken B says:

      I may be way off here but I think it’s simpler. If the effect is felt by an ACA beneficiary as expanding his options in a pleasant way it’s good, if on the other hand it’s felt as burdensome then bad. The distinction isn’t about overalll effects, just how The recipient perceives the incentives facing him. Y and D just use the jargon as labels for this subjective perception of the ACA beneficiary.

  9. Benjamin Cole says:

    I welcome the scrutiny of the ACA, and I think the ACA should be scrapped.

    I would also welcome such scrutiny of the 3.7 million veterans now receiving “disability” payments of $57 billion annually, and both figures are rising rapidly. (BTW, the Pentagon says about 150,000 Americans have been injured in battle since and including Vietnam).

    So 3.7 million Americans receive $1284 a month literally for doing nothing. There is no trust fund for this; it is paid for 100 percent by taxing productive citizens. Pay as you go.

    If you give a man a fish, then you have to give him a fish the next day. If you teach a man to fish….shouldn’t vets take mandatory job training to get off of disability?

    This is what saddens me: many issues that the right-wing will talk about extensively…and the many that they will not talk about….ethanol…$1 trillion a year for DoD, DHS and the VA…extensive NSA spying, in peacetime….the USDA….the vast pink wonderland that is rural America….

  10. John says:

    Let me ask a question: if the ACA serves as a disincentive to work, is the same true of social security and Medicare. If so, are those programs in some sense “disastrous” and should they be eliminated?

    • Bob Murphy says:

      John, I’m not so sure about the impacts of Social Security, since it’s actually regressive. But yes, I definitely think Social Security and Medicare are disastrous programs that make Americans poorer than they otherwise would be.

  11. John says:

    Yeah, I just think, as judges sometimes write, it is simply much too late too argue that social security and Medicare are somehow “disasters.” Before them, tens of millions of the elderly lived in penury and died ugly deaths. That’s why the programs were passed. I have been reading this blog a lot over the months and I’ve tried to understand the views of the people writing here, but at the end of the day I just don’t. I see the point that government programs like social security limit freedom in some ways, and I understand that the government forces us “through violence,” as people sometimes write here, to pay for them. But this world that some here seem to imagine, which must end with the poor, the unlucky, the stupid, dying in the streets, doomed to eat cat food at 65 years old in cold water flats (as they once were), is not a place in which I want to live. I recognize that in some sense that really is a limitation on personal freedom. Intellectually I understand the argument that in a perfectly free world none of us would be forced to help those who, through bad luck, indolence, or foolishness, lose life’s lottery, or who didn’t or couldn’t save enough in old age, or who can’t afford to pay for doctors, etc. But both intellectually and emotionally, I just can’t agree that that perfectly free world, at least as I understand it, would be anything other than a nightmare of human suffering.

    • Dan says:

      “But this world that some here seem to imagine, which must end with the poor, the unlucky, the stupid, dying in the streets, doomed to eat cat food at 65 years old in cold water flats (as they once were), is not a place in which I want to live.”

      Nobody wants to live in a place like that. I find it bizarre when people act like libertarians just don’t care about human suffering, and that’s why we hold the beliefs we do. I became a libertarian largely because I believe it is the way to provide a better life for everyone. I care so much about human suffering that I want to end all wars, stop throwing people in rape cages for victimless crimes, stop dropping bombs on innocent men, women, and children, stop allowing police to brutally murder homeless people and get away with it, stop the government from torturing people, stop indefinite detention, stop taking children from their parents for smoking pot and putting them in the care of someone who ends up killing the child, and on and on. You may not agree with our solutions, just as we don’t agree with yours, but please stop insinuating that we don’t care about human suffering.

    • andrew' says:

      I suspect the benefits of ss are exaggerated.

      The net costs are the below unity multiplier if not sometimes negative. Not to mention the program insolvency- possibly due to incentive effects – evidenced by the need and inability to raise retirement ages.

  12. andrew' says:

    The big red herring involves that the employer side involuntary part-timing and disemployment is going to be worse than people think and probably more than estimated. Did the CBO give us the prior underestimation?

Leave a Reply