Saturday, July 14, 2012

Mankiw and CBO's analysis stacks the deck a little, methinks

He writes:

"To update one of the tables for the next edition of my favorite textbook, I have been looking at the new CBO report on the distribution of income and taxes.  I found the following calculations, based on the numbers in the CBO's Table 7, illuminating.

Because transfer payments are, in effect, the opposite of taxes, it makes sense to look not just at taxes paid, but at taxes paid minus transfers received.  For 2009, the most recent year available, here are taxes less transfers as a percentage of market income (income that households earned from their work and savings):

Bottom quintile: -301 percent
Second quintile: -42 percent
Middle quintile: -5 percent
Fourth quintile: 10 percent
Highest quintile: 22 percent

Top one percent: 28 percent

The negative 301 percent means that a typical family in the bottom quintile receives about $3 in transfer payments for every dollar earned.

The most surprising fact to me was that the effective tax rate is negative for the middle quintile.  According to the CBO data, this number was positive from 1979 (when the data begins) to 2007.  It was negative 0.5 percent in 2008, and negative 5 percent in 2009.  That is, the middle class, having long been a net contributor to the funding of government, is now a net recipient of government largess

The exercise is fine, in my opinion. He's exactly right that payments are like negative taxes. But why focus on transfer payments when assessing net contributions? I don't see the justification for that. There are lots of things the government spends revenue on besides transfer programs that provide real benefits to people across the income distribution. Simply spending on security and the maintenance of property rights confers huge benefits on the upper quintiles, as does the provision of general education and training and a dependable infrastructure. Of course everyone along the income distribution benefits from this, but this also happens to be where the benefits accruing to higher quintiles primarily come from. In certain cases, like security spending and spending on the justice system, the benefits of the spending increase with income. To the extent that the criminal justice system is slanted against low income populations, one could even argue that spending on the justice system is regressive.

So why focus just on direct transfer payments?

I think this sort of analysis is useful for thinking about the direct impact of tax and transfer programs on disposable income, but I don't think it's good for saying things like "the middle class, having long been a net contributor to the funding of government, is now a net recipient of government largess."


  1. Of course most of this is because the largest recipients are the retired and poor. Excluding them would also be illuminative.

  2. Well, "largesse" is a loaded term. Furthermore, identifying the middle income group as middle class is incorrect. The middle class typically earn more than that. "Transfer payments" is also a loaded term, despite being a term of art. From what I hear, a fair number of people in the middle income group are on food stamps. Oh, the transfer payments! Oh, the largesse!

    I wonder, did Mankiw include the tax breaks and loopholes that the rich take advantage of in his negative tax calculations? Did he include the subsidies to agribusiness and Wal-Mart, among others, in the transfer payments?

  3. I don't think things like protection of property rights are beneficial only to the upper quintiles. Military/government spending greatly benefits the middle quintiles as well since many of them work there. Can you also count government employment as something of a transfer payment considering that it generally compensates above market levels?

    1. As for your first point - I obviously agree as I noted in the post. I'm not so sure about that last point. I haven't looked at the question closely but some smart people would challenge the premise of it.

    2. This quote made me think you were implying that upper quintiles benefited disproportionately from spending on defense of property rights: "Simply spending on security and the maintenance of property rights confers huge benefits on the upper quintiles, as does the provision of general education and training and a dependable infrastructure."

      I'm sure you're familiar with Hernando de Soto's work on property rights and how they benefit the poor: Anyway, I guess you agree that the poor derive huge benefits from the protection of property rights, but the wording is just strange.

      I would think that the idea of calling compensation beyond market rates paid by the government a transfer payment would not be terribly uncontroversial. Sure, you could argue that government compensation is *not* above market, but assuming it is I don't see how that would be a controversial position. Do you think those same smart people would find it controversial if government compensation was below market rates (which would also be a transfer payment to all non-government workers)? It could go either way.

    3. I'm not sure why it's strange - doesn't the protection of property rights confer huge benefits on the upper quintiles? Doesn't most of the benefits the upper quintiles come from this sort of stuff? Anyway, we seem on the same page.

    4. I am old enough to remember when gov't jobs were considered to be low paying. What has happened since then, it seems, is that private sector jobs do not pay so well anymore. I suppose that that represents a transfer from private employees to employers.

    5. Direct transfer payments make sense. It's easy to measure and there are no glaring a priori reasons to think that non-direct benefits (overall) vary in proportion to income. To investigate non-direct benefits you'd need a lot of data, most of which probably isn't available. Which in turn means you'd need to develop some kind of model or metric. If you're making a preliminary conjecture or doing a bit a looking around, it makes sense to use direct payments as a proxy.

      And by the way, if anything, I would argue the opposite. That is, low income tends to confer proportionally larger non-direct benefits than higher income. His estimates are arguably conservative.

    6. Low income tends to confer proportionally larger non-direct costs, as well.

  4. Daniel, two words, "Team Republican," where the truth never matters.

  5. I wish Mankiw would shut up about his wretched textbook.

  6. My response to this post may be summarized as "Oh come on".

    I think it's probably true that the justice system spends more on the rich than others, because they are probably involved in more property disputes. But, in many cases things are reversed. Poor people have a much higher probability of being criminals and therefore of costing the state more to try them and house them in prison, and so on.

    But, for many other government services the cost is per person or per household. And for some poorer people are likely to cost more.

    I don't know if rubbish collection is state service where you are, it isn't where I live. Anyway, will it cost significantly more to gather the rubbish of a rich person than a poor person? Probably not, but they will pay more taxes. What about mental health for example, I understand that much of the cost of that is borne by the state in the US. Let's say every person's chance of going mad is equal, in that case the richer sections of society will subsidize the rest. What about the cost of roads? Again richer people probably drive more, but not in direct proportion to their income.

    1. Ya - "oh come on" is about my response to this comment too.

      I'm not challenging the idea that the burden of the state is progressive. Of course it is. I think it should be, too. What I'm concerned with is looking just at direct transfer payment spending and then concluding that the middle quintile is getting more than its giving.

      As for the justice system - you're probably right that poor people have a higher probability of being criminals, but also critical is the fact that they have a higher probability of being arrested and convicted for crimes regardless of whether they are criminals or not. That's a cost the state imposes on them, not a cost that they impose on taxpayers.

      I just want this to be put into perspective. If all you're looking at is cash transfers and taxes, just make claims about the state and disposable income, please. You're not talking about the benefiting from the "largess" of the state. You're talking about one particular component of it.

    2. Mankiw is making a point about who depends upon who, about which classes in society subsidize other classes. You criticize him for making that point only using transfer payments. You point to government expenditures and zoom in on a few expenditures where the benefit to the upper classes is greater than the benefit to the lower classes. But, you ignore those expenditures that don't work in that way. You ignore those that are effectively costs per individual and as such are borne progressively, or where the lower classes cost more per capita.

      You don't know that the expenditures you single out are more important than the type I mention. It could well be that once costs to the state are accounted for more broadly the -5% figure Mankiw gives is more negative, not less.

    3. You seem to think I'm denying the point that the burden of government is progressive. I clearly haven't. And I clarified that that's not what I'm saying just above.

      What Mankiw looks at really speaks to the impact of government on disposable income, not the extent to which people differentially benefit from government. Do you disagree?

    4. If you put it like that you're right. But you original post implied that the overall transfer effect of government would show that the lower quartiles are subsidized less than Mankiw's figures do. That's not certain, a more careful analysis may show that they're subsidized more, we don't know until someone does that analysis.

    5. That's exactly how I worded it in the post!!!!

  7. Good poker players regard money in the pot as belonging to the pot, not to the players who made the bets. Once you put money in the pot, you relinquish ownership. Now, of course, it is not illogical to think that the money in the pot belongs to the bettors, until the hand is over. However, thinking that way leads inexperienced players to make irrational decisions, because they are reluctant to part with "their" money that is in the pot.

    Similarly, taxpayers would do well to relinquish any idea of ownership over the taxes that they have paid, except in their roles as citizens and voters. Taxes paid belong to the gov't. Thinking about it that way will help to avoid irrational decisions, just as in poker.

    Note that thinking of taxes paid as belonging to the gov't eliminates the idea of transfer payments. :)


All anonymous comments will be deleted. Consistent pseudonyms are fine.