The definition of money doesn’t matter (and there are no “wrong” definitions)

In the comment section of a recent post lots of people objected to my definition of money, which is the “monetary base.”  Most did not seem to recall that Larry Summers once made the same argument (that falling interest rates are deflationary), in a paper explaining the Gibson paradox. And he used gold as the medium of account.  BTW, the fact that low interest rates being associated with low prices was considered a “paradox” shows that “reasoning from a price change” was a widespread problem until Larry solved the puzzle.  (credit also to his coauthor Robert Barsky (who probably did most of the work), and a slightly earlier paper by Chi-Wen Jevons Lee and Christopher Petruzzi.)

Of course we all know that monetary economics has regressed in the last 5 years, and based on recent comments I’ve received the Gibson paradox relationship is once again viewed as a “paradox,” not the natural implication (during periods where the supply of money was relatively stable) of the downward sloping demand for money as a function of nominal interest rates.

Another objection was that my monetary base definition of money is weird, and in some sense “wrong.”  If only I understood that demand deposits could also be used as money, I’d see how false my claims really are.  OK, just for today I’ll give you all your definition.  For today M2 is money, and the monetary base is called “SumNerdyProfessor’sObsession.”  Let’s shorten that to the base = SNPO. Now I want you to re-read all the posts I’ve written since February 2009, and replace “money” with “SNPO,” everywhere you see the term ‘money.’  Or at least pretend to.  OK, now all my posts are rewritten. Has anything changed?

Nope, all my arguments are equally valid for changes in the supply and demand for SNPO.

I’m begging everyone—no more complaints that I have the wrong definition for money.  A rose by any other name . . .

PS.  No one seems to research the demand for money anymore, but when I was young it was the most researched topic in all of economics.  There are 100s, maybe 1000s of old empirical studies that support my claim that falling interest rates are deflationary, holding the level of base money SNPO fixed.


Tags:

 
 
 

25 Responses to “The definition of money doesn’t matter (and there are no “wrong” definitions)”

  1. Gravatar of dannyb2b dannyb2b
    22. March 2014 at 06:02

    SSumner

    what do you call the funds in your bank account if not money?

  2. Gravatar of Garrett M Garrett M
    22. March 2014 at 07:25

    The funds in your bank account are a claim on money. You trade in this claim for money when you withdraw currency from the ATM.

  3. Gravatar of Frances Coppola Frances Coppola
    22. March 2014 at 07:33

    I’ve just commented about Summers’ paper on the other post – hadn’t seen this one.

    Summers specifically says Gibson’s correlation cannot be demonstrated in periods of fiat money. It is a phenomenon of the classical gold standard.

    Having said that, if you hold base money fixed, what you have is to all intents and purposes a gold standard – as I said in my first comment on the other post. Under these circumstances low interest rates are indeed deflationary. Though when base money is fixed, deflation can be an indicator of growth. David Beckworth has an interesting analysis of US deflation in the classical gold standard period: http://people.wku.edu/david.beckworth/postbellumdeflation.pdf I don’t think a European analysis from the same period would be quite so positive, though.

    It seems to me that what we are arguing about is whether or not the base is – or can be – fixed in a fiat money regime, not about whether low interest rates are deflationary when the base is fixed.

    And we are also arguing about the definition of money. Under a gold standard, is “money” the paper that people use in everyday transactions, or the gold upon which it is based? By defining “money” as the monetary base in a fiat money system where banks create most of the medium of exchange, you are in effect saying that under a gold standard “money” is gold, not paper.

  4. Gravatar of Michael Byrnes Michael Byrnes
    22. March 2014 at 07:34

    “Falling interest rates are deflationary, holding the level of SNPO fixed”

    One of the many things that is confusing me about this post and its predecessor:

    What, if any, causal inferences you are making here?

    Are you saying that falling rates (if money is money fixed) *cause* deflation? Or just that with money fixed, falling rates are indicative of a deflationary environment?

  5. Gravatar of The Market Fiscalist The Market Fiscalist
    22. March 2014 at 07:37

    “There are 100s, maybe 1000s of old empirical studies that support my claim that falling interest rates are deflationary, holding the level of base money SNPO fixed”

    Would it be more accurate to say that if the monetary base is fixed then interest rates can only fall if either the demand to take out loans falls, or the desire to save increases – both of which will be deflationary ?

  6. Gravatar of Scott Freeland Scott Freeland
    22. March 2014 at 08:00

    You guys are missing the point. Stop discussing what to label as money and just follow the model.

  7. Gravatar of Frances Coppola Frances Coppola
    22. March 2014 at 08:01

    Or rather, being consistent with what I said before, that interest rates are procyclical when base money is fixed – i.e. that low interest rates will be associated with deflation, high ones with inflation. I did not mean to imply that low interest rates cause deflation when the base is fixed.

  8. Gravatar of Frances Coppola Frances Coppola
    22. March 2014 at 08:02

    Scott Freeland

    Surely you don’t expect us to stop thinking and questioning?

  9. Gravatar of ssumner ssumner
    22. March 2014 at 08:06

    Danny, I do call them money, just as I say “Bill Gates has a lot of money” when I mean he has a lot of wealth. Money is slang for wealth. My point is that monetary models are easiest to work with if we specifically define money as the base, and excludes Bill Gates’s “money.” You can include DDs, but it makes things much more complicated, with no benefit.

    Garrett, Yup.

    Frances, You said:

    “Having said that, if you hold base money fixed, what you have is to all intents and purposes a gold standard – as I said in my first comment on the other post. Under these circumstances low interest rates are indeed deflationary.”

    Low interest rates are always deflationary, it’s just that during periods where the money supply changes the deflationary impact of low rates may be offset by the inflationary impact of a rising Ms. That’s Summers’s point.

    You said;

    “It seems to me that what we are arguing about is whether or not the base is – or can be – fixed in a fiat money regime, not about whether low interest rates are deflationary when the base is fixed.”

    No that isn’t the argument. No one in the entire world thinks the base is fixed under fiat money regimes. But it certainly can be fixed, and was fixed from August 2007 to May 2008.

    Your final point misses the point of this post. It doesn’t matter how we define money, it matters how we use our definition. We could abolish the term “money” and just use terms like gold, the base, M1, M2, M3, M4, etc. If we did that it would always be clear what we are assuming. As long as we make clear which definition of money we are using, it doesn’t matter if two people disagree on what “money” as a general category is (or should be considered.) We all agree on what the base is, we all agree on what M1 is, etc. When I say money I mean the base.

    Michael. I am saying that lower interest rates cause lower price levels, holding the base constant. It would be like saying more demand for apples causes higher prices, holding the supply of apples constant. The latter claim obviously does NOT imply that years of high apple consumption are correlated with years of high apple prices. And my claim does not imply that low rates are correlated with low prices (although they sometimes are.)

    Market Fiscalist, That would be the way a highly intelligent Keynesian would explain the causal relationship. Like Larry Summers, I prefer to look at the supply and demand for the medium of account, but if you prefer that Keynesian model, it’s fine.

  10. Gravatar of Frances Coppola Frances Coppola
    22. March 2014 at 09:00

    I don’t think “fixed” is quite the right description:

    http://research.stlouisfed.org/fred2/graph/image.php

  11. Gravatar of Frances Coppola Frances Coppola
    22. March 2014 at 09:05

    Sorry, broken link. This should work:

    http://research.stlouisfed.org/fredgraph.png?g=u1l

  12. Gravatar of Major_Freedom Major_Freedom
    22. March 2014 at 09:11

    Money is a generally accepted medium of exchange.

    Base money counts.

    Demand deposits created through credit expansion counts.

    Physical cash counts.

    This shouldn’t be difficult.

    ————–

    In other news, Dallas Fed President Fisher recently stated that “QE was a massive gift intended to boost wealth.”

    That word “gift” was obvious to economists all along, but some political strategists who are more concerned with maintaining the Fed system rather than with the truth, make the claim that it wasn’t a gift at all because banks gave back treasuries in return.

    Or maybe if 200 million people plus one believed it, then it will make that claim a truth. If it were only 200 million people, then it’s not. This works for whether the Earth if flat as well. If enough people believe it’s flat, then it is.

  13. Gravatar of BC BC
    22. March 2014 at 09:12

    Scott, since you brought up Keynesian models in your response to Market Fiscalist and your prior post was on supply and demand for money base, I have a question about that.

    Suppose we write money demand as Md = P*Y/V = (1/V)/(1/NGDP). So, for a given 1/V, we indeed get a money demand curve like the one you drew in the prior post, with 1/NGDP on the y-axis. Now, suppose money demand shifts to the right, i.e., V decreases. Regardless of the cause of the demand shift (fall in interest rates, financial crisis, etc.), it would appear that the Fed could stabilize NGDP by shifting Ms to the right to match the shift in Md. In fact, it would appear that by targeting NGDP, the Fed could effectively ensure that any shifts in money demand are met by shifts in money supply. At least, that is my understanding of market monetarist views.

    Now, nothing in the above statements would appear at first glance to be affected by interest rates, either r>0 or r=0. My question, though, is what would happen if V depended on Ms? Suppose that when r=0, shifting Ms to the right causes V to decline further so that Ms can never “catch up” to Md. Is that what would give rise to a Keynesian liquidity trap? If so, is it a purely empirical question as to whether this can happen or is there a theoretical reason why Ms must eventually be able to meet Md? Thanks.

  14. Gravatar of Vaidas Urba Vaidas Urba
    22. March 2014 at 09:43

    Scott,

    Suppose that M0 is fixed, and there is an increase in M3, and there is a reduction in interest rates.

    In the 1/p – M0 diagram from your previous post, S does not change. Reduction in interest rates is reflected as a shift in D. Increase in M3 is reflected as an opposite shift in D. What is the net direction of a shift in D? We don’t know.

    You wrote:
    “Between August 2007 and May 2008 there was no change in the monetary base, and yet interest rates fell sharply. Not surprisingly NGDP growth slowed and we tipped into recession.”

    This sounds like reasoning from a price change. Between August 2007 and December 2007, it looks like that the shift in D caused by lower interest rates was smaller than the opposite shift in D caused by other factors, such as expansion in broad monetary aggregates. After December 2007, it was the opposite.

  15. Gravatar of ssumner ssumner
    22. March 2014 at 13:16

    Frances, Yes, not precisely fixed, but there was essentially no increase in the base from August 2007 to May 2008. In the meantime interest rates fell sharply. So that’s roughly what I am talking about.

    BC, You asked:

    “Is that what would give rise to a Keynesian liquidity trap? If so, is it a purely empirical question as to whether this can happen or is there a theoretical reason why Ms must eventually be able to meet Md? Thanks.”

    Yes, that would be a liquidity trap. Is it purely a empirical question? I’m not quite sure. I wonder what others would say. Suppose my counterargument was that the central bank could then buy up the planet’s entire stock of financial wealth with no inflation. Is that thought experiment an empirical or theoretical argument against liquidity traps? It seems theoretical to me.

    Vaidas, On your first point, I entirely agree. As always in economics it’s a ceteris paribus argument. Indeed in late 2007 and early 2008 NGDP didn’t actually fall, it just rose at a lower rate, so those other factors must have been operating, just as you hypothesized.

    On your second point, I wasn’t reasoning from a price change, just citing that period as an example of what my thought experiment might look like in the real world.

  16. Gravatar of John carney John carney
    22. March 2014 at 13:46

    If a non-policy market rate for loans is falling, that must mean demand for loans has fallen or supply has increased, no?

    If it’s supply that has pushed down the interest rate, that means people are already withdrawing funds from present consumption. Prices are already falling in that scenario. The interest rate isn’t deflationary; it’s a sign of a deflation already under way.

    If demand for loans is falling, the situation is more complex. Perhaps business are not expanding because they expect lower rates of return. Perhaps they anticipate falling prices so they are forestalling expansion. In either case, deflation is already under way.

    Aren’t you getting causation backward?

    This is why I asked you why rates were falling in your example.

  17. Gravatar of James in London James in London
    22. March 2014 at 14:00

    Perhaps we need a term for people who get hung up in the definition of money and miss the whole point of market monetarism.

    Definitionists? Pedants?

  18. Gravatar of Vaidas Urba Vaidas Urba
    22. March 2014 at 15:16

    Scott,

    The diagram is useful, but we should always check that we have a pure interest rate shock. Maybe it the Fed who caused both lower rates and higher M2. My point was that we should not reason from the lower rates, we should reason from the underlying shock that caused lower rates. That underlying shock has reduced M0 demand between August 2007 and December 2007.

  19. Gravatar of Vaidas Urba Vaidas Urba
    22. March 2014 at 16:06

    Scott,

    Additional comment, as I suddenly remembered the arguments which I used in similar discussion we had a year or two ago. The problem with August 2007 and December 2007 period is that the Fed did two things. The Fed lowered interest rates (increasing demand for M0), and the Fed eased discount rate policy (reducing demand for M0). Taking everything into account, the monetary policy reduced demand for M0 during the period, which combined with the stable M0 implied growing NGDP during the period.

    So yes, it is not a problem of reasoning from a price change. The problem with the statement below is different: “Between August 2007 and May 2008 there was no change in the monetary base, and yet interest rates fell sharply. Not surprisingly NGDP growth slowed and we tipped into recession.” The actual problem with this statement is that it takes only a half of policy decisions into account, and ignores the other half.

  20. Gravatar of John Carney John Carney
    22. March 2014 at 18:43

    By the way Scott, in the spirit of not arguing about definitions, I offer you this:

    http://www.cnbc.com/id/101327578

    It shows that what happened in 2007-2008 was not just falling rates and a flat base but a complete collapse of transactional securities (high quality liquid assets that can be used to finance or collateralize purchases of financial assets).

    This collapse mattered a lot since these things did perform a lot of obey like functions. Once they collapsed, this drove up demand for dollars and Treasuries, pulling down interest rates even as many bond prices collapsed.

  21. Gravatar of Daniel Daniel
    23. March 2014 at 03:19

    John Carney,

    In the spirit of not missing the point, I ask you this – SO WHAT ? Why do you keep dragging finance into this ?

  22. Gravatar of ssumner ssumner
    23. March 2014 at 05:12

    John, Someone made a similar argument in the other post. You are trying to translate the finding into Keynesian language. If you reach a point where you say “well of course that would happen” then you are probably on the right track. After all, unless the Keynesian model is wrong you’d have to be able to translate any monetarist argument into Keynesianism, and vice versa. For instance when the Keynesians talk about fiscal stimulus, the monetarists translate that as “government borrowing boosts interest rates which lowers real money demand which is inflationary.”

    Keep in mind that I am arguing that any fall in interest rates, for any reason other than monetary policy, is deflationary. That includes less demand for loanable funds and more supply of loanable funds. And note that those two shifts have different effects on the quantity of loanable funds.

    That means that an expansionary impact from monetary policy lowering rates must come from the increased M, not the falling i.

    Vaidas, Yes, but I still think that’s just a “ceteris paribus” problem. Suppose I argued that an increase in the demand for potatoes pushed up potato prices. It might be true that the demand for potatoes rose when people felt poorer, as during a depression, and the effect of the depression on potato prices overwhelmed the microeconomic effect. I was starting with a basic observation, which most people don’t understand, that falling interest rates are deflationary, ceteris paribus, as they boost demand for base money. Then we have a lot of empirical support that this effect often overwhelms the effects pushing in opposite directions (interest rates are highly procyclical.)

    The fact that there are sometimes other policies going on, like the discount rate change you mentioned, is no reason not to explain to people that the primary effect of lower interest rates is deflationary, other things equal.

    John #2, See my response to Vaidas.

  23. Gravatar of Vaidas Urba Vaidas Urba
    23. March 2014 at 08:07

    Scott,

    Some observations:

    - The best description of 2H 2007 would be the following one. The Fed has reduced the interest rates. This was inflationary, as lower rates increased M0 more than they increased demand for M0. The Fed took another inflationary step – it changed the discount policy. This inflationary step has reduced M0. As a result of these two inflationary policies, the net change in M0 was close to zero.

    - “I was starting with a basic observation, which most people don’t understand, that falling interest rates are deflationary, ceteris paribus, as they boost demand for base money.” – this is just a description of a microeconomic mechanism. It is very useful to understand this mechanism, but this is mechanism is just one of the mechanisms that operate when the Fed reduces the interest rate target. And this mechanism is not relevant when central banks reduce policy rates by cutting the interest on reserves.

    - “Then we have a lot of empirical support that this effect often overwhelms the effects pushing in opposite directions (interest rates are highly procyclical.)” Interest rates are highly procyclical when this mechanism does not operate at all. When central banks change policy rates by changing interest on reserves, interest rates are procyclical, and M0 is countercyclical.

    - i is a price (a price of a bond), so one has to be careful and identify shocks that caused i to change, otherwise one is runing a risk of reasoning from a price change.

  24. Gravatar of ssumner ssumner
    23. March 2014 at 09:09

    Vaidas, Yes, interest rates are a price, but they are not the price of money. Hence when interest rates change the demand for money shifts. That’s the standard approach in any discussion of the S&D for any good. If the price of money were to change you’d simply move along the demand curve. But interest rates are a shift variable. If I’d made inferences about the bond market then yes, I would have been reasoning from a price change.

    I’d say the interest fell due to the housing bubble bursting. Because the Fed kept the supply of money stable this caused NGDP to fall. However there were other factors that also shifted the demand for money, which meant NGDP did not fall, but the growth rate slowed sharply.

    Of course I agree that a cut in IOR is totally different, as it decreases the demand for the MOA, but that’s not applicable to this case.

  25. Gravatar of Vaidas Urba Vaidas Urba
    23. March 2014 at 10:05

    Scott,
    OK, so this is not reasoning from price change. But ceteris paribus assumption and partial equilibrium reasoning are risky and you have to be very careful when doing the analysis in practice.

    I agree with the analysis beginning with January 2008, and disagree with it in 2H 2007.

Leave a Reply