Two types of “wheelbarrows full of currency”

In a previous post I tried to translate Paul Krugman’s critique of Steve Williamson from the language of Keynesianism (interest rates) into the language of monetarism (base money.)  The basic insight was that when one thinks of wheelbarrows full of cash one thinks of two very different situations; banks in Japan (and more recently the US) engorged with reserves, and of course the German/Zimbabwe hyperinflations. And these two examples (which are polar opposites) illustrate how careful one must be in drawing causal connections between changes in the monetary base and changes in the price level.  I thought that was Krugman’s point, but given my weak grasp of theory I had a lot of doubt.  Now that Krugman himself (with David Andolfatto) have switched into the language of monetarism, I’m much more confident my previous post was addressing his point, in an oblique way.

But I’d still like confirmation from someone with a higher IQ than me!  Noah??

Here’s Krugman (and Andolfatto):

So, here’s how Andolfatto describes Williamson’s point:

Evidently, one of the effects of QE (in the model) is to increase the real stock of currency held by the private sector, and agents require an increase in currency’s rate of return (a fall in the inflation rate) to induce them to hold more currency. (Remember that the results are all contingent on the way monetary and fiscal policy are modeled.)

OK, so “agents require” a fall in the inflation rate to induce them to hold more currency. How does this requirement translate into an incentive for producers of goods and services “” remember, we’re talking about stuff going on in the real economy “” to raise prices less or cut them? Don’t retreat behind a screen of math “” tell me a story.

PS.  I don’t have any views on Williamson’s model as my math is so weak it would take me a month to have an intelligent opinion.  I disagree with his empirical claim (as does David Beckworth), but that’s a different issue.

PPS.  I do find the debate entertaining.  Kocherlakota’s faux pas a few years back was the initial skirmish; this is the full blown battle.  Kocherlakota himself seems to have switched sides.  Then there are people like Tyler Cowen and Noah Smith, who think Williamson’s being criticized unfairly in theoretical grounds, but have some doubts about his empirical claims (at least that’s how I read Tyler and Noah.)  Sorry to be such a wimp on this, but I’d rather not strongly defend a position unless I have a high level of confidence.

HT:  Mark Sadowski



13 Responses to “Two types of “wheelbarrows full of currency””

  1. Gravatar of benjamin cole benjamin cole
    2. December 2013 at 17:19

    scott Sumner—I take a friendly jab at you and George Selgin and discuss the Minneapolis Massacre at Marcus Nunes’ blog, Historinhas

  2. Gravatar of benjamin cole benjamin cole
    2. December 2013 at 17:28

    Beckworth has excellent post on Williamson…and a reminder just how fragile models are, even simple models…a switch from headline to core PCE can undercut supposed correlations….

  3. Gravatar of dtoh dtoh
    2. December 2013 at 17:42

    You need to distinguish between MB being held as an MOE and money being held as a store of value (SOV).

    The non-banking sector holds the amount of MOE it needs for transactions. Full stop. There is a close correlation between the amount of MOE and NGDP with some slight variations depending on the convenience yield. (Or not so slight in the case of Zimbabwe.)

    When the convenience yield is low or zero, the banking sector will hold large amounts of money as an SOV (ER) if the CB allows or encourages it. The amount of ER has no relationship to NGDP (unless the CB is signalling that they intend to cause the banking sector to reduce ER by exchanging it for liabilities of the non-banking sector or vice versa).

    (Japan is a slightly different case because of a more recent history of and willingness by the private sector to hold large amounts of cash for tax evasion and other institutional reasons.)

    Scott, if you accepted the fact that monetary policy effects NGDP primarily through the mechanism of changing the amount of exchange by the private sector of financial assets for goods and services, then this would all be crystal clear.

  4. Gravatar of Mark A. Sadowski Mark A. Sadowski
    2. December 2013 at 17:45

    benjamin cole,
    If you read my comment at Beckworth’s post, you’ll note that unlike Beckworth, who looked at core PCEPI, my results were for headline PCEPI. So what Williamson claims about QE and inflation, is not empirically true regardless of the measure of inflation used.

  5. Gravatar of Daniel J Daniel J
    2. December 2013 at 18:04

    Scott I think most people that read Steve’s post had the same difficulties but attempted to decipher it with not so good consequences. At least you’re being honest about it.

  6. Gravatar of dtoh dtoh
    2. December 2013 at 18:43

    I think you should have entitled the post “Two Types of Money”

  7. Gravatar of dirk dirk
    2. December 2013 at 19:19

    I don’t understand why the example in this MR post is “deflationary”, as Tyler claims:

    Can someone here explain? I asked for help four times in the comments over there, so that’s why I’m here. This example strikes me as inflationary not deflationary, as it lowers the demand for cash, but I’m not an economist and realize that a lot of these things are counterintuitive.

  8. Gravatar of benjamin cole benjamin cole
    3. December 2013 at 03:59

    Mark S.
    As usual, excellent work by you. Why hasn’t the Fed contacted you and offered you a visiting scholar slot? See my post at Marcus Nunes…

  9. Gravatar of ssumner ssumner
    3. December 2013 at 05:33

    Ben, I agree about the Beckworth post.

    dtoh, I don’t see how that conflicts with anything I said here, or in the other post I link to. I agree that base money is held for many different purposes. ERs have less impact on NGDP, although of course they may affect expectations.

    dirk, I’ll take a look.

  10. Gravatar of jknarr jknarr
    3. December 2013 at 09:25

    I think that this QE-NGDP issue screams out for some formal treatment of debt, i.e. anti-AD.

    Williamson should be commended for being one of the few that treats debt seriously in an AD context. (Though he’s wrong to treat base money reserves and treasury debt as equivalent; and he may be right about deflation, but for the wrong reasons.)

    All debt in the US (3.5x NGDP, not just treasury paper matters) will ultimately needs to be paid back in base money. Hence, debt is also a demand sink for base money (not just NGDP). Supply-and-demand conditions for debt are the interest rate. Low rates means low demand for debt, which in our banking system, means lower base-to-M123 multipliers.

    (Nick’s thing about the car also totally ignores the role of debt. His accelerator is “money supply” (mostly bank liability-based aside from M0), and his speedometer is the price of debt. So yes, pushing the speedometer lower will cause the accelerator to press down (the Q of debt to go up). It’s not a good analogy.)

    I think Williamson may be right for the wrong reasons. Debt is deflationary — try being heavily indebted and see what it does to your AD. QE preserves debt, and creates more room for its formation — e.g. “the treasury needs to issue more debt because of scarce assets!”

    More debt = more deflation, more demand for base money. QE, low rates, sluggish NGDP are the symptoms of high preexisting leverage. Insofar as reserve creation allows for more debt formation, it is long run deflationary.

    The only way to exit QE, low rates, high debt, and sluggish NGDP is deleveraging. This can be via deflationary default, or with base money production — but not with reserves, which can only collateralize debt formation.

    The only near-permanent base money asset that offsets debt liabilities and deleverages is currency. Hence currency = debt, ultimately. That’s your hyperinflation. Or, sit back and accept reserve-based-QE, low rates, high debt, and sluggish NGDP foevah.

  11. Gravatar of ssumner ssumner
    3. December 2013 at 11:07

    jknarr, You said;

    “I think Williamson may be right for the wrong reasons. Debt is deflationary “” try being heavily indebted and see what it does to your AD.”

    This is classic fallacy of composition.

  12. Gravatar of jknarr jknarr
    3. December 2013 at 11:42

    Scott, yes I know, I similarly thought that as I wrote, but I kept it because I’m talking about preexisting debt, which is also deflationary in aggregate!

    i.e. yes marginal increases in savings is bad if everybody does it at once; yes marginal increases in dissavings all at once can boost NGDP.

    But we’re talking more about a preexisting stock of debt across nearly all economic sectors that serves to dampen monetary policy and AD in the here and now — the temporal NGDP boost from past dissaving has run its course.

    Now, the debt remains, and we are at zero rates/low demand as the result of past dissaving. Bottom line, preexisting debt affects the demand for base money and the efficacy of Fed policy.

    More debt means more pent-up future deflation. All debt does is pull demand into the present at the expense of the future. The Fed will ultimately accomodate these debt-fueled demands for base money — first with reserves, second with currency.

    All the reserve-based QE so far has really achieved is keeping debt-default-deflation off the books for a while: the debt remains and must one day be liquidated with deflation or inflated with currency.

  13. Gravatar of ssumner ssumner
    3. December 2013 at 16:04

    jknarr, If I am to make sense of a debt story I have to work it through monetary policy. There may be a connection via interest rate targeting at the zero bound, but that’s the analysis I can follow.

Leave a Reply