Potpourri
Yikes I’ve been traveling lately and am way behind with some of these…
==> Niels van der Linden replies to my comment on his critique of Rothbard.
==> Richard Ebeling (my teacher in undergrad) reflects on the Great Society.
==> Megan McArdle on health insurance bailout. Joe Salerno provides more details on this interesting story at the Mises blog.
==> Somebody writing for Jim Cramer’s (?) website says I am a leading singer in the Austrian doomsday chorus. Bob Roddis, if you’re reading, please save this one for when the market crashes.
==> A reader (sorry I forget who at this point) sends me this discussion of outrageous police deception in manipulating confessions from people. Seriously, these are eye-opening.
==> For those who pine for a revenue-neutral carbon tax reform, look at what’s actually happening in California.
==> I discuss the minimum wage and other issues with Tom Woods.
==> I discuss the world’s 85 richest people with Scott Horton.
==> I am not being sarcastic, I feel bad for Tyler Cowen. On the one hand, he is constantly criticized for being a “sell out” by one group of libertarians, while in the comments at his blog people routinely denounce his free-market dogmatism (seriously). He is well-versed in all kinds of economic theory, and yet has to spend time arguing that demand curves slope downward. No wonder he has an eating disorder.
==> I may have lost an inflation bet, but at least I never lost a radiation bet. Stephen Hawking better hope Brad DeLong doesn’t find out about this…
So, about the whole “tax interaction effect” thing, I’ve been thinking some more and I have a pretty clear example of why the thinking is flawed, that “concentrating a new tax on some ‘bad’ could be inefficient because it requires more distortive taxes”.
Let’s say that we have a situation in which a water resource is unowned, and people just take whatever fish they want from it. Classic tragedy of the commons situation (not far from that alleged in AGW/carbon emissions).
Someone proposes auctioning off the fishing rights and using the proceeds to reduce income taxes.
I’m pretty sure the libertarian reaction would be not just “yes”, but “heck yes”: more resources out of the government’s control, remove a tragedy of the commons, and reduce a damaging tax. Win, win, win.
But the tax interaction effect argument would say that, no, now fish have an actual extraction price (for the rights, not for the fishing rod, etc). That’s functionally a tax that’s “narrowly concentrated” on fish consumers, and thus *extremely* distortive, relative to the fractions of a percent that could be shaved off income taxes.
Yeah, it’s a stupid argument, but isomorphic to the one made against a carbon tax shift. (Though this is orthogonal to your main point in the link which is that we won’t get the tax-cut side of the “shift”.)
Your analogy looks to be quite a good one.
However, if I understood it, Bobs argument was never about the wisdom of a carbon tax, it was against the idea of a “win-win” situation, where we get increased economic activity by swapping taxes. In your example it would be like saying the new arrangement would not only preserve fish stocks, but would also be a boost to the economy through reduced income tax. The argument that the “fish tax” plus reduced income tax would actually boost the economy is unlikely because of the TIE. It may well still be a good idea to have the fish tax, but not because of the win-win. Is this a reasonable representation?
But that TIE argument would be wrong. Because someone now owns the capital value of the fish source, they have the incentive to balance fishing now against fishing later, and thereby prevent overdepletion.
So whatever reduction in consumption follows from people having to pay for fishing rights, would be (more than) offset by the gain in consumption in later periods … and that’s even before accounting for the efficiency gain from lower income taxes.
(Even in the edge case where time/fish preference is extremely high, and depletion *is* the efficient choice, this would, at worst, be a wash.)
Likewise, one could argue, the increased availability of “atmospheric resources” in later period from reduced CO2 emission now would — assuming optimal taxation — more than offset the harms from reduced fossil fuel usage.
Exactly. The argument is not against carbon tax win, but against a win-win. The “fish tax” would have benefits in the future by avoiding over-depletion. The carbon tax would have benefits in the future by reduced environmental damage (or increased atmospheric resources). Both of these are paid for by costs today. In the fish case, fish will cost more today. In the carbon case, energy will cost more today.
Some people have argued that there would be benefits today also, since the carbon tax (or fish tax) could reduce income tax. I think the TIE shows why this may not be the case. The TIE is not saying that Pigovian taxes can’t have benefits in the longer term, but that they probably will not have benefits today.
I am going to run with this argument for a bit to see if it makes sense. If our goal is maximum efficiency of the economy, the TIE reduces the optimum level of tax on the “bad”. In the fish case, if the water were owned by an individual, he would not be interested in overall efficiency of the economy, but in his own maximum gain. He would set the price of permits to maximise his own yield. If fishing an extra fish costs an extra amount X due to reduced fish stocks in the future, he will set the permit cost at X. Effectively he would set the price at the “social cost” of fishing. However, if we were to do this as a tax, the TIE shows that in a revenue neutral case, the transfer of some tax from a more efficient income tax to a less efficient (more concentrated) fish tax will have a negative effect on the overall economy. This means that for overall efficiency maximisation the fish tax should be set lower than X. This will result in fish costing a bit more in the future than would have been the case with the social cost tax, but this will be compensated by greater efficiency in the economy generally.
So in the carbon tax case, the most efficient tax level would be less than the social cost. This will result in more costs due to carbon in the atmosphere, but this will be compensated by greater efficiency in the economy overall.
As far as I know, there is no way to determine accurately how much lower than the social cost it should be set for maximum efficiency.
This makes sense to me, but I may have got the wrong end of the stick.
Silas this is intriguing. I’m swamped with stuff right now; if I don’t get back to you on this, please remind me in a month or two.
“Niels van der Linden replies to my comment on his critique of Rothbard.”
FTA:
I think the best way of looking at this particular question is to consider the following ideas:
1. If a good is rising in value as a medium of exchange, then strictly speaking that means more people would be holding that good. So far, Murphy, Rothbard, and van der Linden would agree.
2. What does this imply about the total quantity of money? According to Rothbard and Murphy, the total quantity would tend to increase. According to van der Linden, not necessarily. I think van der Linden is wrong, but I think there is an even better way of showing why, which is slightly different from Rothbard’s “marketability” idea and Murphy’s “money as a consumer good” idea. That is, just consider what happens to the supply of ANYTHING over time if there is a continuous positive demand for it. If there is a continuous positive demand for cars, then as technoligical advances and capital accumulation increases the productivity of labor, the supply of cars will tend to increase. Consider gold. If gold were money, and there was a continuous positive demand to hold gold, then just like with cars, as technology and general productivity improve, so will the rate of producing gold improve. No one individual has to decide how much total supply of cars or gold there has to be. The supply of gold or cars specifically will increase because of the demand, and hence investment, technology, and capital accumulation.
At 10:30-11:00 or so in your talk with Tom you talk about there being negative employment effect in studies looking at a adjacent states. I don’t think this is true – the state level studies (typically – I think Sabia has one out with an adjacent state look) do not look at adjacent states, they just have state fixed effects. So it’s not like the textbook story guys are doing good counterfactual matches and then the other guys are throwing in funny time period controls. They generally aren’t doing that (which is bad because we think there is spatial heterogeneity), and then even weak controls for spatial heterogeneity (for example, differential time trends by state or region) make those negative effects go away.
Strong controls for spatial heterogeneity (actually matching the counties) has no disemployment effects either.
I also found it funny that Tom said only Austrians don’t think you need government to address externalities. He should google Coase 🙂
Emphasis on “I don’t think” – let me know if I’m wrong. That’s certainly not what DLR do when they reproduce the NW results. As far as the NW side of the literature I’ve been focusing on Meer and West. But I don’t think it’s the case.
I always chuckle a little when I see people talking like they can actually observe minimum wage laws reducing or having no effect or boosting employment.
Even if the government enforced a $1 million an hour minimum wage rate starting tomorrow, and alongside this we observed skyrocketing unemployment, we still could not conclude that A caused B just because we saw A correlated with B. For in the method of empiricism, there is assumed a 5% chance of being wrong, such as omitting variables unrelated to minimum wage that might have had an effect on unemployment.
It is not good enough to say “It’s so “likely”, so “probable”, that it is “reasonable” to conclude that the minimum wage “almost assuredly” caused the unemployment. This is just rhetoric designed to cover up one’s admission that we cannot conclude from A correlated with B that A caused B.
The actual ways of knowing whether or not minimum wage laws case unemployment is to explicate on what minimum wage laws ARE, and what wage payments ARE, and through this understanding, deduce on praxeological grounds what the relationship between A and B really are. That is how you do economics.
Inferring from adjacent state statistics, time period control statistics, spatial heterogeniety statistics, is a deeply misguided data mining exercise that can only serve to keep the knowledge uncertain, which of course encourages the researcher to fill in the blanks with his own a priori beliefs.
In DK’s case, he wants to have his cake and eat it too. His prior writings ooze a priori theory in the face of data mining. He wants for minimum wage to remain intact, and for it to have only positive or no effect on employment. So massage the equations and keep things just uncertain enough, to be able to present ones’ pro-minimum wage theory as scientific, while at the same time attacking anti-minimum wage theorists as dogmatic. The irony, ’tis thick enough to stop a bullet.
Yep. Mises made this clear in Human Action. Observing what happened in the past and using it to make conclusions about the future is history, not economics.
You are not using the term “data mining” as it’s typically used, MF.
Anyway, think what you will. The rest of us will be happy to continue the discussion without you.
You are not using the term “data mining” as it’s typically used, MF.
Perhaps, but from the perspective of someone who knows that economics is based on a priori reasoning, it is not unreasonable to argue that ALL studying of history that is intended to find economic truths, is “data mining.”
Anyway, think what you will. The rest of us will be happy to continue the discussion without you.
Maybe with enough years of ignoring something, it will eventually become false.
“The actual ways of knowing whether or not minimum wage laws ca[u]se unemployment is to explicate on what minimum wage laws ARE, and what wage payments ARE, and through this understanding, deduce on praxeological grounds what the relationship between A and B really are. That is how you do economics.”
The problem here is that you have no way of knowing whether your explication of what minimum wage and wage payments are is complete. If you have missed some aspect of either, then your conclusion will not accord to what happens in practice. Of course, you will have a perfectly logical and accurate description of what would happen IF your analysis were accurate and complete.
So what did he miss then? What part of his analysis is inaccurate or incomplete?
The minimum wage crusaders never address this. They just say “studies prove it isn’t true!” I have yet to hear anyone offer a serious critique based on a priori reasoning that would explain WHY demand curves slope downward for everything EXCEPT low-skilled labor…
I don’t know what he missed – he didn’t explain what the result of the explication was. My point is that you can never know if you have missed something.
Harold:
“The problem here is that you have no way of knowing whether your explication of what minimum wage and wage payments are is complete. If you have missed some aspect of either, then your conclusion will not accord to what happens in practice. Of course, you will have a perfectly logical and accurate description of what would happen IF your analysis were accurate and complete.”
We do not require “complete” knowledge before we can know certain relationships are true.
At any rate, if you say it is a problem that we can’t know minimum wage fully nor wage payments fully, then any testable model that utilizes such variables, would suffer from the same problem. So you can’t claim your method is any better than mine.
“My point is that you can never know if you have missed something.”
So by that logic, you can never know if you missed something in the course of you making that very conclusion you just made about never knowing.
Harold, it’s no use attacking reason and logic. You’d just end up undercutting your own arguments.
I know the geometry analogy is done to death, but I’m gonna make it anyway.
“We can’t know if the Pythagorean theorem is correct. We might have missed something. The only way to be certain of it is to measure a bunch of right triangles and perform a regression analysis, controlling for what color paper the triangle is drawn on and other significant variables, and if 99% of triangles support it, then we can declare it correct.”
Don’t confuse logic with effects in the real world. Within logic we can know things because we define things. I can claim that it is logically impossible to know if we have a complete description without evidence. That is why I say you can have a perfectly logical and accurate description of what would happen IF your analysis [of what minimum wage laws and wage payments actually are] were accurate and complete. It is only if you then want to say this is what would actually happen if you,
say, increase minimum wage that you have a problem.
Another triangle test. All angles add up to 180 degrees. We can prove this, measure a million triangles and it works every time. Then we draw a triangle on a curved surface. Oh – it looks like we missed something.
“At any rate, if you say it is a problem that we can’t know minimum wage fully nor wage payments fully, then any testable model that utilizes such variables, would suffer from the same problem. So you can’t claim your method is any better than mine.”
I can, because if my method reveals a discrepancy I know I have to go back and look for something else. It can reveal that I have missed something.
Another thought on the triangles. For theoretical triangles we can prove Pythagoras. If we have a real, physical triangle, Pythagoras’ theorem will not be enough for us to know if Pythagoras applies. Perhaps the angle is not quite 90°.
For theoretical economic agents we can prove minimum wage reduces employment using a particular economic theory. If we want to know if that will actually happen, we have to know that the real agents are as described by the theory. This is like checking if our triangle actually is a right angle triangle.
DK wrote:
At 10:30-11:00 or so in your talk with Tom you talk about there being negative employment effect in studies looking at a adjacent states…
Well, I was just being sloppy there. If you listen to the way I introduced that snippet, you’ll hear that the context was the DLR 2010 approach that looked at pairs of adjacent counties that lied in different states. And my understanding is, if they don’t include geographical-specific time trends, then even regressions just on this subset (with just fixed effects) shows a negative effect.
That’s cool – I’m as much writing this out to confirm that I’m understanding.
The fixed effects does show a negative effect, but they have NOT introduced the pair fixed effects at that point. When the pair fixed effects are included it’s not negative.
So in other words, the specification where they get the negative effect is done with the counties, but you are not restricting the comparison to only the adjacent counties yet. You are still just comparing all treatment counties to all control counties UNTIL you add the county pair fixed effect, at which point you are only using the within-pair variation to identify the treatment effect (because all the between pair variation is soaked up by the pair fixed effects that are added.
In other words, there is no “matching” until the very last specification.
Daniel,
OK I totally get what you are claiming. I think I had that impression from the Schmitt CEPR summary of Dube et al. 2010, but I just re-read it and Schmitt isn’t saying what I thought he said.
I will double-check Dube et al. before making this point again, to be sure I’m accurate.
“I may have lost an inflation bet, but at least I never lost a radiation bet. Stephen Hawking better hope Brad DeLong doesn’t find out about this…”
It would be funny if Hawking did what DeLong yammered on about a while back. Hawking would have to say:
“I have been totally wrong, about everything. I am closing down my publishing for five years to avoid misleading readers while I intellectually retool. You will find me sitting at the feet of Dr.Preskill, chanting ‘om mani padme hum’ until I achieve enlightenment.”
Of course, MF, if DeLong wanted to he would say, “Right you idiots, Hawking changed his mind. That’s what scientists do.” So we probably shouldn’t push the analogy too far…
Except your theory of the relationship between quantity of money, money holding, and price inflation, wasn’t refuted by the fact that price inflation did not pain out as you predicted.
You were just wrong in your prediction regarding the extent to which people held cash. Your theory of the relationship between quantity of money, money holding, and price inflation, was fully consistent with what happened.
You yourself changed your mind about the extent to which the demand for money can rise. You even admitted as much when you said “I didn’t know just how much people would hold onto their money” (paraphrasing).
DeLong doesn’t have you nailed AT ALL.
In other words, if DeLong can speak like that about you given that you changed your mind about the extent of money holding times, then to be consistent he’d have to do the same thing with Hawking.
Is there a less clear writer than Tyler Cowen?
Major Freedom, put your hand down.
Blame yourself for your reading comprehension problem.
Ken B needs to read “I heart violence” between the lines before a political philosophy argument “makes sense” to him.
It’s why DK’s and LK’s posts often make him giddy. Makes him feel less guilty about his psychopathy.
Don’t forget Keshav.
“I may have lost an inflation bet, but at least I never lost a radiation bet. Stephen Hawking better hope Brad DeLong doesn’t find out about this…”
This was an awesome line btw.
Not that it is any consolation but there was inflation. It went into lowered velocity. Whoever thibkks they won because the fed can’t or won’t create labor inflation may have believed there wasn’t inflation during the housing bubble because the fed refused to measure it.