More on the Paradox of Thrift

Here's Matt Yglesias, making the argument that I've been trying to cash out for the past week:

Now we’ve entered “paradox of thrift” territory. People are saving more. And the increased saving isn’t being cycled back into the economy as new investment. In part, that’s because of problems in the financial system. But in part, it’s because with short-term demand slumping so much, there’s not a lot of worthwhile investing to be doing. The economy needs someone to decide to borrow some money and start a new firm that employs these newly unemployed people. But with the volume of consumption going down so rapidly, nobody’s really in the mood to start a new business. And existing businesses are busy scaling back production, not interested in borrowing money to ramp it up. The result of this is an overall fall in the average level of income. And that means that even with the share of income being saved going up, the actual level of savings can be going down and we can truly end up in the toilet.

Tyler Cowen thinks Matt's explanation is mostly right, although he says that it's wrong to think of Americans saving more, since they weren't really saving all that much to start with. According to Cowen, it's more accurate to say that the paradox could result instead in rising levels of debt to GDP. Specifically, if people begin to "save" by paying back credit card debts, then that's okay if they're paying those debts to relatively healthy banks that can turn around and lend the money back. But if those payments are mostly going to zombie banks, then the bank might well just hold on to the cash. And that can create problems.

I'm inclined to think that this is pretty much the argument I've been making. Though, obviously, if I could make it as well as Matt and/or Tyler, then I'd be either a famous blogger or an econ professor at George Mason.

Share this

Neither one really makes the argument

They're both giving part of the whole picture and trading on their reputations as reputable economists that what they've left out doesn't impact the conclusions one is likely to draw.

If you say, "I feel itchy", then the obvious solution is "well then, scratch." But depending on the larger picture, that may not be the best course of action. If you feel itchy because of poison ivy or because of some other skin condition, there may be good reason to refrain from scratching, because your skin can only take so much abrasion.

Similarly, if you give part of the whole picture, this partial picture, even if true, may suggest a useless or counterproductive response.

I thought this discussion was more complete.

I'm not sure...

...that I see how Waldman and Cowen are saying anything all that different. Cowen says the problem currently is that people may (though we don't know this for sure yet) be using a lot of their excess disposable income to pay down debts at zombie banks who then sit on the money rather than reinvesting it. Isn't that just a specific instance of Waldman's broken financial system?

It's also worth noting that the Waldman post was written at the end of July 2008. Things have worsened significantly since then. Or, more to the point, a lot of banks that still looked reasonably solvent in July (I'm looking at you BOA and Citigroup) we now know to be in zombie territory. Not coincidentally, BOA and Citigroup are (or were as of the end of 2007, anyway), the two largest holders of credit card portfolios.

Again, none of this is to say that one need draw the same conclusions from admitting that we're in paradox of thrift territory. Yglesias and Kurgman, for example, support a giant stimulus package. OTOH, as Jim Henley pointed out this morning, Cowen, along with Alex Tabarrok and Jim Hamilton (Keynsians none of them), think that short-term aid to states as well as targeted infrastructure spending might be the solution.

I think that you're conflating two different arguments. Or, at least, that you're conflating two different arguments:

  1. Keynes is right that modern economies can sometimes enter into paradox of thrift territory (though not for the reasons Keynes actually thought).
  2. If we accept that Keynes is right about the paradox of thrift, then we must also accept Keynes' own proposed solution to the paradox.

I have been arguing that accepting (1) does not logically entail accepting (2). One can accept that Keynes' description is correct while denying that his prescription is correct. Instead, I've been arguing for something more like targeted infrastructure spending (for which I think there's a case to be made on good free market grounds anyway). You seem (still) to want to insist on some sort of logical connection between (1) and (2). I'm still not seeing it. And, at the risk of appealing to authority, plenty of people who are better at economics than either of us aren't seeing that connection, either.

Plenty of economists

It is boring to point out the 200 economists that Cato rounded up to come out against the stimulus.

"Notwithstanding reports that all economists are Keynesians and that we all support a big increase in the burden of government, we the undersigned do not believe that more government spending is a way to improve economic performance."

As far as I'm concerned, their authority cancels out the authority of economists in favor of a stimulus, throwing us back on actual arguments.

Meanwhile, there is that other problem, the spectre of a

Catastrophic Collapse, in which a loss of confidence by investors in U.S. government debt leads to a total collapse in the U.S. financial system, with economic activity contracting by 90 percent or more.

which possibility argues against massive spending.

I have been arguing that accepting (1) does not logically entail accepting (2).

(2) clarifies the meaning of (1). If you recommend additional spending, this clarifies that you think the underlying diagnosis, rather than merely a symptom of a deeper diagnosis, is a lack of spending. If, however, a lack of spending is merely a superficial symptom of a far more important problem, then additional spending can exacerbate the underlying problem (e.g. by sending misleading price signals). This includes both additional spending on digging holes and filling them in, and also additional spending on "targeted infrastructure projects".

Analogy: if someone feels hot and your recommendation is to lower the room temperature, then this clarifies that your diagnosis is not malaria.

Deja vu...

It's the same argument all over again. I'm really quite puzzled. There are lots and lots of examples of people correctly describing problems but incorrectly prescribing solutions. For example, Marx is right that religion is the opiate of the masses. But his solution is disastrously wrong. Accepting the correctness of Marx's description doesn't entail that I believe (a) that the reason on which he based that conclusion is correct or (b) that his solution to the problem is correct.

which possibility argues against massive spending.

Right, but if you actually go back and read my various posts and my various comments, you find that I am not endorsing massive spending. Again, you seem to assume that believing that the paradox of thrift exists entails supporting the stimulus bill as written. But there simply is no logical connection between these two things. Which brings us to...

If you recommend additional spending, this clarifies that you think the underlying diagnosis, rather than merely a symptom of a deeper diagnosis, is a lack of spending.

...which is a textbook example of affirming the consequent.

I give up

...which is a textbook example of affirming the consequent.

No, it isn't. Affirmation of the consequent is a kind of argument, but you did not quote my argument (which immediately followed), only the conclusion. A conclusion quoted separately from its argument is not the argument and therefore is no fallacy, even if the argument is. But lately you seem to be misconstruing my arguments in trivializing ways, so I'll give up.

Look, I'm not intending to

Look, I'm not intending to trivialize your argument. And your reply is a bit nitpicky; the fact that I didn't blockquote your entire argument is hardly a basis for concluding that I somehow don't understand what an argument is or that a fallacious one involves both premises and a conclusion.

But the fact remains that your argument is a bad one. You are still assuming that there is somehow a logical connection between accepting a description that Keynes might support and accepting Keynes' reasons for offering that description. Moreover, you assume that accepting a solution that Keynes might (at least partly) support logically entails accepting all the other solutions that Keynes might have supported. But in the spirit of constructive dialog (which, I think, this has mostly been, despite our disagreements), I'll give it another shot.

If you recommend additional spending, this clarifies that you think the underlying diagnosis, rather than merely a symptom of a deeper diagnosis, is a lack of spending. If, however, a lack of spending is merely a superficial symptom of a far more important problem, then additional spending can exacerbate the underlying problem (e.g. by sending misleading price signals). This includes both additional spending on digging holes and filling them in, and also additional spending on "targeted infrastructure projects".

Your argument, as I see it, is this:

  1. If you think that Keynesianism is the correct way to understand macroeconomics, then you will support increasing government spending to jump-start the economy.
  2. You support increasing government spending to jump-start the economy.
  3. Therefore you have to be committed to Keynesianism as the correct way to understand macroeconomics.

Yes, I know that's simplified. We could break some of those into subpremises. But the overall logic of your last comment seems to boil down to that argument. But that, however, is an instance of affirming the consequent. I can support increasing government spending for all sorts of other reasons. To avoid affirming the consequent, you need to replace (1) with:

    1' You will support increasing government spending to jump-start the economy if and only if you think that Keynesianism is the correct way to understand macroeconomics.

Now you have a conclusion that follows logically from your premises. But it comes at the cost of introducing a premise, namely 1', that is just flatly false. It's totally unclear to me why one can't consistently hold, for example, that government spending on infrastructure will stimulate economic growth while at the same time rejecting the claim that just any old government spending will jump-start growth. Russell Roberts, for example, argues for a similar claim in the Globe.

Maybe, though, I've seriously misunderstood what you're trying to argue. Did you have a different premise (1) in mind? Maybe you could construct a premise 1'' that shows me how you derive (3) from something plus premise (2).

No

You are still assuming that there is somehow a logical connection between accepting a description that Keynes might support and accepting Keynes' reasons for offering that description.

No, you have started interpreting everything I might say through the lens of a peripheral remark I made a while ago about how you might most clearly go about arguing your case.

My main point, which I've been trying to make and which you keep shoving into a little square hole, is that economic errors are often caused by not looking at enough of a large enough picture. I've been trying to get you to focus on a larger picture and you're just not doing it. It's not a question of me confusing your argument with Keynes's argument - you presented multiple versions of your argument and I read them, and I find them incomplete. I'm done now. Maybe some other time.