What Evidence Bad Arguments?

Megan McArdle Wednesday:

Chait, and others writing in this vein, refute the strongest claims of the supply-side movement: that tax cuts produce astonishing growth, or that cutting taxes can increase tax revenue. Then they imply that they have thereby refuted all the economic claims in favor of tax cuts, which they haven’t, not even close. … it’s not some sort of weird, uniquely awful Republican behaviour to sell your policies using dubious economic claims. … Politicians often assemble policies for a variety of reasons, and then sell them using the least plausible, but most appealing, rationale.

Any social worker, for example, will tell you that a core of their clients have no reasonable chance of getting off support. … But almost no social worker ever says, "We need welfare benefits because these people are too screwed up to hold a job", because Americans do not care to give money to people whom they perceive as not trying. Whether this is appalling dishonesty, or merely putting your best foot forward, depends much on how you feel about the underlying program.

If there were no relation between the quality of a policy conclusion and of the public arguments offered in its favor, then there would be simply no point in considering public arguments when evaluating a policy.  Yet Megan spends most of her time critiquing public arguments offered for various policies. 

So I think Megan must admit that the weakness of some arguments for tax cuts was in fact a bad sign about the wisdom of those cuts.  Her point, I think, is that it wasn’t an overwhelming bad sign; Megan sees other good signs that could outweigh this bad sign. 

GD Star Rating
Tagged as:
Trackback URL:
  • I see two questions:

    1) Is this a special evil that Republicans are prone to, and therefore should merit special moral outrage from me that would force me to vote for Democrats?

    2) Are these bad arguments the dominant force driving either the crafting or passage of these policies?

    I see the answers to both as no. My objection is not that Chait takes on a bad argument, which he ought; it is that he presents that bad argument as a core of pulsating evil at the heart of the Republican party that ought to thoroughly discredit them. That is very silly.

  • < ?xml version="1.0" standalone="yes"?>
    < !DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">

    I don’t agree that “Megan must admit that the weakness of some arguments for tax cuts was in fact a bad sign about the wisdom of those cuts;” she has, I think, effectively admitted that it’s a bad sign about the wisdom of the voters, and about political honesty (on both sides, she says, with some evidence.)

    It’s possible for a politician to choose policies rationally, and then rationally choose irrational arguments (i.e., lie) to get them accepted (because he has rejected “The Myth of the Rational Voter”). It may not be happening, but it’s possible. Certainly I’ve never seen a whole lot of honesty on either side of any elections…I thought I did, back when Carter ran for election. I was young.

  • Eric Falkenstein

    I’m against affirmative action. So are Neo-Nazis, who are wrong to assert Aryan racial supremacy. By Robin’s logic, I must therefore admit that the weakness of some arguments against affirmative action are a bad sign about the wisdom of those arguments, and revise my beliefs appropriately. Every bad idea that supports my conclusions weakens conclusions, irrespective as to whether I believed, or even knew about them!

    Robin is making the assumption that some argument supporting A is wrong, so A should move closer to the opposing argument B. But Megan says B is filled with just as many overbroad arguments, and she didn’t assume A’s bad arguments anyway.

  • Eric, the post did say “public arguments offered in its favor”; the existence of Neo-Nazi arguments itself isn’t evidence, but if they were the arguments that were used the most by affirmative action detractors, then that would be indirect evidence that no better arguments existed.

  • Eric Falkenstein

    Steven, ‘public arguments’ in favor of your conclusions, if faulty, should cause you to adjust your beliefs in your conclusion, even if you are using different assumptions, or a different model, to arrive at your conclusion? The mere existence of bad ‘public’ arguments is sufficient for adjustment?

    I would think that bad arguments in favor of things you believe could be relevant to adjusting your beliefs, but they could be irrelevant. They don’t help, certainly, but idiot reasoning supporting my sundry beliefs are to my mind irrelevant, not disconcerting.

  • I guess that whether there should be a serious adjustment depends on the issue. If I just encountered a controversy for the first time, I would be very interested to learn that the most high-profile arguments for one side were wrong. But if I had reason to believe most participants were operating on bad assumptions or using a bad model compared to my own, the quality of their arguments may no longer contain much information value. You could probably analyze this with causal diagrams, like arguments from authority.

  • Eric Falkenstein

    Steven. I think you and Robin are looking at two countervailing beliefs, and evaluating them like a judge hearing two biased advocates. To the extent one side uses bad arguments, that counts against them, the more so the more they relied on it. This reasoning is not applicable to evaluating beliefs that one has great familiarity with, if those ‘bad’ beliefs are not a key part of one’s reasoning.

  • Right, I think we agree, then. Public argument quality is a useful indicator *on average*, but it may be useless in many particular cases — specifically, the ones where you know enough to rely on your own understanding rather than on some estimate of the credibility of the people on either “side”.

  • If I think (as I do, as I think everyone posting does) that persons X (Bush, Giuliani et al.) offer invalid arguments for Y, then this affects my opinion of X: it excludes all models in which they are simultaneously (a) well-informed about Y, (b) sufficiently intelligent to understand the information, and (c) honest about their information. Right?

    The trouble is, I didn’t believe (c) to begin with; I believe that even if X understood Y perfectly, and knew exactly how to explain the validity/invalidity of the various arguments for and against Y, they would place almost no value on doing so. They want to win. They accept the irrational views of the marginal voter as providing hurdles which is is their job to overcome — not “overcoming bias” in the sense of this blog, but “overcoming bias” in the sense of getting over all the hurdles. Winning.

    So, in response to Steven’s thought that it might be “indirect evidence that no better arguments existed”, that could be true, but “better” means “more convincing to the marginal voter.” And they (Bush etc.) might be right…these invalid arguments may be the most effective when placed in an ordinary speech.

    Unfortunately, this doesn’t help me decide about Y — the actual subject matter — at all. For that, I read Mankiw, DeLong, Cowen, Rodrik, Kling, etc, when I have time — and mostly end up deciding, as Megan does, that it’s the spending that matters. Mostly.

  • There are usually more positions than A & B. Pretending that they’re all that exist is preying on a common human dialectical bias -where’s the internal transparency in this thread about that?

    Also, there may be a “Hannity and Colmes” effect: where a dialectic is presented and in the pagentry one side seems to be deliberately presented weaker than the other.

  • “The world’s stupidest man may say the sun is shining, but that doesn’t make it dark out.”

    If you look around the Web, the main argument you find for the Singularity is Moore’s Law for computing power, which, as I’ll be noting tomorrow, is mostly irrelevant. Anyone who sets out to refute the assertion that computing power grows along a predictably smooth exponential curve is not refuting a strawman. It is an argument put forth by the field’s most famous popular advocates. But it is not the strongest argument. People do not know what is strong. They do not repeat what is strong. Bad arguments are not evidence until you hear them from a field’s most technical advocates. The popular forums are not evidence. Never mind what a politician says about tax cuts – let me hear from a small-“l” libertarian economist, and if those arguments are bad, I may give up.

    A technical critique of a popular argument is not evidence that no good arguments exist, because the intervening popularity filter may have filtered out all the good stuff.

  • Tom Breton

    (Arguendo I’ll assume Megan’s account of the situation is accurate and
    that Chait is indeed arguing something like “They made a bad argument”
    -> some implicit logic -> “Their position is wrong”)

    “Evidence by bad arguments” is a case of argument by failed search.
    It’s equivalent to saying that one’s efforts constitute a failed
    search for a credible argument for the opposing position, therefore
    they haven’t got one, or probably haven’t got one.

    In this case, the conditions for a successful argument by failed
    search aren’t well met.

    Did he exhaust the cases directly? No, he only rebutted two
    arguments. More than that were raised. Anyways, it would be quite
    ambitious to hope to find and refute every pro-lower-taxes argument.

    Did he exhaust the cases a fortiori, then? That is, by refuting the
    best arguments, so that we could reasonably assume the rest, being
    worse, were flawed too?

    He would also have had to argue that these cases were the strongest
    arguments, unless it’s obvious. More than that, they must not merely
    be a little stronger on balance, because this appraisal may easily
    change, especially when good counterarguments are raised. They must
    collectively dominate the others. Ie, more technically, there must be
    no reasonable way to weight factors that results in a dominating set
    that is not a subset of this set.

    But he raised no argument and it’s not obvious. On the contrary, it
    looks like he selected the easiest targets.

    Did he probabilistically characterize the cases by sampling, then?

    For that, he would have to show that the population of arguments he is
    selecting from fairly represents the
    position. It can’t be, for instance, a few good
    arguments mixed with many strawmen.

    He would also have to argue that he sampled fairly from that
    population. Alternatively, he could argue that he sampled it
    charitably: better than fairly, making judgement calls in the
    opponent’s favor.

    Again, he raises no argument, it’s not obvious, and his selection
    clearly was not random.

    That’s the sort of thing he would have to do to make a reasonable
    argument by failed search.

  • J Thomas

    If the prominent arguments for X are idiotic, it doesn’t at all mean that X is wrong. It does mean that the people making those arguments for X are evil or stupid.

    You can excuse them for their evil. “They just want to win, and lying is the way to win.” That argument can be extended — in the long run the way to win is to keep people from being educated, but instead work to give them prejudices that will make your lies easier to believe. This isn’t a slippery slope, it’s evil from the first step.

    And when it’s a question of approving of the actions of the people who’ve shown they’re evil or stupid — I notice that the US public was ready to let Bush start wars and bomb foreigners and all. But I also notice that as soon as he wanted to take their Social Security money with flimy excuses all of a sudden they didn’t trust him. All of a sudden he looked evil or stupid to them. They cared. There could be valid arguments to reform SS, but when the question was whether to let Bush do it….

    If you know about advanced arguments that work, while the popular arguments are trash, you can do a public service by finding ways to explain the advanced arguments in popular language. You’ll lose some concision and clarity because the advanced terminology was designed for the purpose and popular language is not. But it’s *worth doing*.

    One possible format is to start by saying that often reality (and advanced explanations for reality) is counterintuitive. Maybe give an example or two of counterintuitive things. “Well, you know, sometimes things are just backward from what you’d expect. Like, I met a man from somalia who’d never seen a toilet before, and he was amazed that when you pour a bucket of water into it that it flushes instead of overflowing. But that’s how they work. And there’s this idea that when everybody is greedy and grasping and they try to grab as much as they can for themselves, that everything will work out for the best and we’ll all be as rich as we can be. I’m not certain that’s right, but it certainly doesn’t *sound* right. There are lots of things like that.”

    And then make the argument in simple language. “These guys say that if you lower taxes, everybody will have more money and they’ll spend more money and there will be more work for everybody and then there will be more to tax and so the government will take in just as much as before. That reminds me of a story, this guy was going to have a mink farm, he’d raise minks and skin them and sell the fur for lots of money. And he figured he didn’t need to spend anything on food for the minks because he could feed them the skinned minks and that would be enough. Somehow this tax thing reminds me of that story.

    “Well, we could cut taxes. But if we do cut taxes don’t we need to cut government spending too? If the government keeps buying stuff without actually paying for it, the stuff the government uses is stuff that you can’t buy. All the *things* the government sucks up are things that somebody had to work for that nobody gets to enjoy. All the lead that goes into bullets for the army is lead that doesn’t go for bullets for hunters, and so on. That’s whether the government gets the money from taxes or not. Don’t look at the money shell game, look at the real stuff. If we want to be richer it isn’t enough to cut taxes, we have to get the government to use less stuff. Buy less stuff from contractors and fire government workers.”

    And end up with the argument from authority. “The scientists who study this stuff won’t say how big the government ought to be. But they say we don’t take enough tax that reducing the tax will help things much, and they say when you cut taxes you need to cut government spending too. They say it and it’s plain common sense too.”

    This is too long. And it might be too plain. You have to fit it to the audience. But it’s a whole lot better if you can make true arguments — even if you have to simplify them a bit and you lose professional clarity — than if you settle for lies that sound good.

  • I’m a bit confused. To suggest that tax cuts produce estonishing growth is clearly hyperbole, but saying they may increase revenue is unequivocally true. So long as a tax cut increases growth, it will inevitably produce a revenues greater than the alternative at some point in the future and thereafter.

  • I also don’t remember hearing the argument that tax cuts would create “estonishing growth”. That sounds like something Chait made up.

    He seems to be more guitly than his target. They certainly exaggerated, mostly by ommision, as an increased revenue stream is atleast 30 years out for an increase of growth of .25% or less for a tax cut of 2%, but I don’t remeber people claiming that increased revenues would happen instantly. (That’s what some people might have assumed, and they were happy to let them believe it).

  • J Thomas

    Aaron, which claims are we discussing? The original Laffer curve claims looked to me to assume equilibrium conditions.

    It sounds like you’re talking about a different argument entirely. Anything we did that increased GDP growth by, say, 1% per year would eventually have a gigantic effect on the economy. But why would you expect tax cuts to do that?

    Try three approaches.

    A: We collect a larger amount of taxes, and as a result private citizens have less money to spend so demand is down, and they have less money to invest so supply is down.

    B: We collect less money in taxes. Instead the goverment borrows the money as bonds. As a result private citizens have less money to invest, and they hire fewer people and/or pay them less, and demand is down. (In theory the government will someday pay that money back, but in practice instead of paying off the interest they borrow more money to pay the interest, and keep reborrowing indefinitely.)

    C: We collect no money in taxes at all, we completely abolish taxes and finance the government entirely through bonds.

    Which of these approaches would you expect to appreciably increase GDP, or increase the rate that GDP increases?

    If the government not only reduced taxes but also cut spending, that might increase GDP. If we, say, fired 5% of our civil servants and 5% of the military, if we cut back all government services and military spending by 5%, then that would provide a pool of skilled unemployed job-seekers for private companies to employ, and it would release a tremendous amount of resources for private industries to convert to consumer goods. Mountains of paper, mountains of steel, mountains of depleted uranium, hoards of MREs, etc. All the resources the government doesn’t consume are available for consumers to consume, once they’ve been transformed to something consumers want.

    Without reduced government spending, there will still be effects from the government getting its money from bonds instead of taxes. Lots of subtle effects. Would you expect those effects to amount to a significant increase in GDP?

    But then, this isn’t the only argument in favor of reduced taxes. A more important argument goes like this: If I promise to reduce *your* taxes then *you* will have more money. Somebody *else* can pay for the government. I want *your* vote.

    Bribing voters is a time-honored practice, though it isn’t exactly *honored*.

    I’m still mostly arguing this like a layman, though I’ve started using more classy words. Is my argument wrong? Is it too hard to tell what the argument is, without the specialised language?

  • Guys, this isn’t the place to argue about tax policy.

  • J Thomas

    Robin, I took a freshman math class where we were given some axioms and a collection of theorems to prove, one third of which were false. The teacher said, “If you’re going to learn to use symbolic logic it’s better to actually practice it than talk about theory.” I’m not sure that approach is completely right for all cases, but it did work in the cases I saw.

    The claim is that there are at least two levels of discourse, one where people actually use logic and a second where the herd of citizens listens to sleazy rhetoric to get wrongly persuaded.

    Possible responses to this are to argue only on the abstruse level and lose politically, build cunning lies on the sleazy level and perhaps win politically, ignore the problem, or look for a way to effectively tell truth among the sleaze.

    We are practicing the second and the fourth possibilities, studying the methods by example. Which of us is doing what is left as an exercise for the reader.

    Should we do that somewhere else?

  • Apologies in advance Robin,

    J Thomas, no I wouldn’t disagree much with what you wrote. We are probably pretty much in equilibrium, or will soon hit it if the interest rate we carry our debt at rises much. If the interest rate were to rise significantly, then we’d probably be better off raising taxes. But if we do get more economic growth from a tax cut, it makes sense.

  • J Thomas

    Aaron, I believe you’re mostly disagreeing with what I wrote.

    Here’s an analogy.

    It’s like, you have a young donkey, and you use it to carry hay to the market. The bigger the donkey grows the more hay it can carry. But to grow it has to eat. And the more of the hay the donkey eats the less it brings to market today. If you starve the donkey it can’t carry as much even today, and if you try to make it carry too heavy a load it will limp along or even refuse to move. Arguing about taxes versus bonds is like thinking about how to balance the load, ignoring the issue of how big the load is or how much of it the donkey gets to eat. Maybe the heavy load isn’t being balanced the very best it could be, but that isn’t the main issue.

    The big issue is government spending, not taxes versus bonds. Cutting taxes without cutting spending is just bribing the particular people whose taxes were cut.

    Am i saying it clearly enough? Am I wrong?

  • Clear enough and wrong (for the most part). Correct spending is the real issue, yes. But given a level of spending, if you have a cheaper source of capital, you use that. If it were true that the interest was greater than the gain in economic growth, there’d simply be no economic growth since the interest would be higher and is subtracted from GDP.

    You’ve confused sources and types of food for your donkey with balancing the load.

    Again Robin is right, wrong place. There is probably a better forum at MarginalRevolution, Econlog, or Megan McArdle’s. E-mail me if you wish to continue.

  • Sorry, thought my e-mail was on my website.