Andy Harless on monetary policy

I’ve been asked to comment on a recent Andy Harless post that denied the existence of monetary policy:

So there you have it: there is no such thing as monetary policy. There is “central bank directed stabilization policy,” and, for convenience, you can refer to that as monetary policy if you want. If so, recognize that you are using the term loosely, and let’s not get into arguments about whether some particular Fed action is “really” monetary policy. None of it is really monetary policy.

I agree that when one takes a close look at monetary policy, things are never quite what they seem.  And that it’s hard to draw a bright line that separates monetary policy from other policies.  But I still think one can identify two basic types of monetary policy:

1.  Policies that affect the supply (or quantity) of the medium of account. 

In the US, the Fed creates two assets that serve as the medium of account; currency and bank deposits at the Fed.  This aggregate is referred to as the monetary base.  Open market operations are the primary method by which the supply of base money is altered, although discount loans are also used.  Thus the base is increased when the central bank buys an asset, or when the central bank lends base money to others.

2.  Policies that affect the demand for the medium of account.

The central bank can affect the demand for base money in numerous ways:

a.  Reserve requirement changes.

b.  Changes in the interest rate on reserves (and perhaps currency when we fully shift to electronic money.)

c.  Changes in the nominal policy target.  Thus a higher price level target, a higher NGDP target, or a higher foreign exchange rate peg would tend to reduce the real demand for base money, by creating higher inflation expectations.  Indeed any action that increases the future expected supply of base money will tend to reduce the current real demand for base money.

To summarize, an expansionary monetary policy is any action or statement by the Fed that increases the supply of base money, or reduces the real demand for base money.

I see base money as being special because it is a medium of account.  Thus the price level can be modeled in terms of changes in the value of “money.”  T-bills are different.  An increase in the supply of T-bills might increase the price level, or it might reduce the nominal price of T-bills.  Money is special; by convention its nominal price is equal to one.  Fiscal policy has little impact on the price level.  When Reagan generated huge deficits via tax cuts and a military buildup, and Volcker fought back with tight money, we all know who won.  And the Fed also came out ahead when LBJ tried a tight fiscal policy in 1968-69, but monetary policy stayed easy.  That’s why monetary policy is important, and worthy of special attention.


Tags:

 
 
 

42 Responses to “Andy Harless on monetary policy”

  1. Gravatar of Andy Harless Andy Harless
    11. December 2010 at 19:39

    I would challenge the premise that “base money” is equivalent to “the medium of account.” Why not say that cash is the medium of account and bank reserves are an asset whose value is derived from that of cash? If you’re going to include bank reserves as part of the medium of account, then why not include deposits: they have the same property of being freely convertible to cash; the only difference is who does the converting and who is allowed to hold them. Or you could include T-bills and money market funds in the medium of account. Or you could exclude all of these things, even paper money, and include only coins minted by the Treasury. How is the definition not arbitrary?

    Before I wrote the last paragraph, my inclination was to take cash as the medium of account, but now that I think of it, the value of paper money is arguably derived from that of coins. Paper money just consists of notes redeemable for coins. They say “Federal Reserve Note” on them, and the word “note” indicates that they are, theoretically, a promise to pay and not an actual payment — although they also say “legal tender,” which means one is obliged to accept them in lieu of actual payment.

    In the end, I’m still inclined to include paper money in the medium of account — but not interest bearing deposits, even if those deposits are held at the Fed. Bank reserves are “as good as cash” because there is no question about the Fed’s ability or willingness to redeem them in cash. But they also bear an interest rate that the Fed can change at will. In that sense, they’re like bills with a maturity of one second (or one day?) that automatically roll over, but that don’t guarantee a yield for the rollover. The medium of account includes the principal portion but not the interest portion. That seems arbitrary to me, and it seems again that, in their essential nature, they are more like T-bills (which you apparently don’t consider to be part of the medium of account) than like cash.

  2. Gravatar of Andy Harless Andy Harless
    11. December 2010 at 20:01

    What I forgot to say in my last comment (though I already said it, more or less, in the original post, but I think it bears repeating) is that if you take cash as the medium of account — which I think is a reasonable definition — then the Fed’s policy is not to control the supply of the medium of account but to accommodate all demand. So the “supply side” monetary policy is a degenerate and essentially unvarying case, not really relevant to “monetary policy” as we normally speak of it.

    Now in this post you include “demand side” monetary policy as an additional case. The problem I have with this nomenclature is that all sorts of things change the demand for the medium of account, and it’s arbitrary to say that those things are monetary policy when the Fed does them but not when someone else does. In fact, the Fed could affect the demand for the medium of account by making statements about fiscal policy. If Ben Bernanke encourages Congress to pursue a fiscal stimulus and the market believes his statement will have an impact on Congress’ decisions, that is literally a “statement by the Fed that…reduces the real demand for base money,” but it’s hard to believe anyone would consider that statement to be monetary policy. And it’s hard to see why it should be considered monetary policy when Ben Bernanke does it but not when some influential academic economist does it.

  3. Gravatar of Jon Jon
    11. December 2010 at 21:53

    Andy: bank reserves are ultimately the medium through which your cheques and drafts are settled. The banks don’t carry wads of cash back and forth. So if you accept that cash is the medium of exchange, you ought to except that bank reserves are also the medium of exchange because ultimately the chequeing system is only viable if settlement is possible.

  4. Gravatar of Lorenzo from Oz Lorenzo from Oz
    11. December 2010 at 22:53

    In using the term ‘medium of account’ are people saying that the unit of account is also used as a medium of exchange? (Which need not be the case, though there will be some rate of exchange of one into the other if it is not the case: think cigarettes in PoW camps or tinned mackerel in contemporary US prisons were smoking is banned.) If not, what is meant?

  5. Gravatar of Bill Woolsey Bill Woolsey
    12. December 2010 at 05:30

    The unit of account is the word used to quote prices, make contracts, and keep accounts. It is a word like dollar, yen, or euro. The medium of account is what is used to define the unit of account. Going from a silver to a gold standard changes the medium of account, with the unit of account staying the same (dollar, pound, etc.)

    Generally, some monetary asset serves as medium of account. The issue here is that in the U.S. some monetary assets serve as medium of account and others don’t.

    Private bank liabilities don’t serve as medium of account. Banks are contractually required by redeemability to keep the value of their deposits at par, but it is certainly possible, and has occured sometimes, that deposits in insolvent (or even illiquid) banks trade at a discount. This doesn’t make sense for the medium of account. It’s price is fixed by definition.

    You can imagine that deposits at the central bank are just like deposits at private banks, and there is a contract (redeemability) that keeps the deposits pegged at par to hand-to-hand currency, which is the real medium of account, but I don’t think this makes much sense. The central bank can issue hand-to-hand currency at will. Redeemability provides no contractual constraint on the creation of reserve balances.

    That is why monetary theorists (and maybe especially quasi-monetarists) aggregate hand-to-hand currency and reserve balances at the central bank.

    This is a bit obscure, but when private banks can issue their own hand to hand currency, treating bank deposits as if their are redeemable in the banks own hand to hand currency doesn’t make much sense. Both currency and deposits are redeemabily in some other base money–perhaps central bank deposits or currency or else gold or silver.

    Also a bit obscure is considering systems without central bank currency. If all the currency is privately issued and cleared like checks with deposits at the central bank, or else, there is no currency, and all payments are made by electronic or check transfer of deposits, then deposits at the central bank still serve as medium of account.

    And, of course, if we turn that around, and the central bank doesn’t have deposits, and banks settle up net clearings with government issued paper money (perhaps held on deposit at private clearing banks,) then the currency remains the medium of account.

  6. Gravatar of Bill Woolsey Bill Woolsey
    12. December 2010 at 05:34

    “Changes in the interest rate on reserves (and perhaps currency when we fully shift to electronic money.)”

    Interest on currency when we fully shift to electronic money?

    Oddly put. I would say that when we fully shift to electronic money there is _no_ currency. All payments are made by electronic transfer of deposits.

    What would you count as “currency” on which interest could be paid in a fully electronic payments system? How is it different from a deposit?

  7. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    12. December 2010 at 05:55

    “I see base money as being special because it is a medium of account. Thus the price level can be modeled in terms of changes in the value of “money.” T-bills are different. An increase in the supply of T-bills might increase the price level, or it might reduce the nominal price of T-bills. Money is special; by convention its nominal price is equal to one. ”

    The price level can be modeled in terms of par value of T-bills – you’ll get the equivalent results.

    T-bills are not so different from money, as changes in the nominal price of T-bills is mirrored in changes of scarcity and convenience yield of money, and for any theory about base money a mirror-image T-bill theory can be built, the predictions will be equivalent.

  8. Gravatar of Andy Harless Andy Harless
    12. December 2010 at 05:58

    Jon, the phrase used in the post was “medium of account” rather than “medium of exchange.” It’s not necessarily the medium actually used to do transactions, but the medium in terms of which the values of transactions are defined. Banks don’t use cash to settle transactions (nor do non-financial entities use bank reserves), but the values of those transactions are defined in terms of cash.

  9. Gravatar of Bill Woolsey Bill Woolsey
    12. December 2010 at 05:59

    Mr. Money Demand:

    Any good can serve as numeraire in a GE model.

    And that includes the par value of T-bills.

    What is the market process by which an excess demand for T-bills leads to a lower price level?

    Not, find the price level where the supply and demand for goods and services and T-bills and everything else is in equilibrium. But rather, describe exactly how an excess demand for T-bills results in people lowering the prices they charge for various goods and services.

  10. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    12. December 2010 at 07:14

    Bill Woolsey,

    We should focus on those goods that provide at least some services in facilitating exchange, as excess demand for liquidity services of one such good spills over to all others. Broadest monetary aggregates usually attempt to include all such goods.

    Look at the Lehman crisis timeline. The price of T-bills soared on September 18. Why did this happen? It happened because there was a flight to money. Demand for T-bills increased because the supply of shadow banking system money has crashed. On the other hand, the price of the 10 year note has dropped on September 18. The rest is history.

  11. Gravatar of scott sumner scott sumner
    12. December 2010 at 09:10

    Andy, Reserves are a medium of account, as their price is fixed. T-bills are not a media of account.

    I’m actually sympathetic to the idea of using cash held by the public (i.e currency and coins) as the measure of money, now that reserves pay interest. But I don’t see the distinction between currency and coins. Both are fiat money, now that coins are no longer silver.

    I don’t have a good answer about demand deposits. I suppose I mentioned the base because we were discussing monetary policy–and the Fed can directly determine the supply of base money. But I suppose any definition is a bit arbitrary.

    Andy, You said:

    “What I forgot to say in my last comment (though I already said it, more or less, in the original post, but I think it bears repeating) is that if you take cash as the medium of account “” which I think is a reasonable definition “” then the Fed’s policy is not to control the supply of the medium of account but to accommodate all demand. So the “supply side” monetary policy is a degenerate and essentially unvarying case, not really relevant to “monetary policy” as we normally speak of it.”

    I don’t quite agree with this. I don’t think Fed changes in S-T interest rates have much direct effect on the economy, Rather, when the Fed cuts rates it is signaling to markets that it intends to increase the base (including currency) and that this will lower short term rates. In addition, the increase in the base and currency will raise the future expected level on NGDP (through the excess cash balance effect) and that explains how the policy shift affects current AD.

    Regarding other factors that effect the S&D for base money, I entirely agree in a qualitative sense. But the effects are small in comparison to the Fed’s tools and more importantly, CAN, SHOULD BE, AND OFTEN ARE OFFSET BY FED COUNTERMEASURES TO KEEP NGDP GROWTH ON TARGET.

    Lornezo, The medium of account is the object that embodies the unit of account. It is X, as in “the dollar is defined as one unit of X.” Bill explains it better than I can.

    Bill, That’s an excellent summary of the issues. I would just add one point. In my work on the depression I treated gold and currency as dual media of accounts. I primarily focused on the gold market as the source of shocks, but noted that when the price of gold changed in 1933, currency continued to be a unit of account whereas gold did not. Then in 1934 both went back to being media of account.

    Bill#2, Currency would be prepaid cash cards, redeemable in central bank reserves, whose balances rise over time at the rate of interest on central bank reserves. But perhaps that’s not really the same.

    123, I don’t agree about T-bills as the par value is not equal to the market value. To be a medium of account the market price must be fixed. Otherwise all of monetary theory breaks down. The value of money is no longer 1/P.

    BTW, I am not arguing that changes in the price of T-bills don’t have an important impact on the demand for money–obviously it can, at low interest rates.

  12. Gravatar of Andy Harless Andy Harless
    12. December 2010 at 09:43

    Scott: “when the Fed cuts rates it is signaling to markets that it intends to increase the base

    That was the case before the Fed started paying interest on reserves. Going forward, I expect that Fed will change the IOR rate in parallel with its funds rate target (and that it will continue to accommodate cash demand), so increasing the base will no longer be effective as a means of reducing interest rates. A cut in in the target rate will signal an increase in deposits via the market mechanism (as banks respond to an immediate reduction in the IOR rate) rather than as a signal of the Fed’s subsequent intentions about base money.

    other factors that effect the S&D for base money…the effects are small in comparison to the Fed’s tools

    I don’t agree with this. Especially if base money pays interest, then fiscal actions have a dramatic effect on the demand, because government bonds are close substitutes for base money. Not perfect substitutes, but close. If the government makes more purchases and issues more T-bills to finance them, the demand for base money will go down significantly because banks that were holding base money (and to some extent their customers that were holding interest-bearing deposits) will switch to holding the newly-issued T-bills.

    It’s true that (at least under normal conditions) the Fed can offset fiscal policy-induced changes in the demand for money, but the converse is also true: fiscal policy can offset the impact of Fed actions. Whoever happens to move last gets to decide. The Fed is likely to move last because the institutional process is smoother, but I don’t see this as a reason to separate the Fed’s actions conceptually as “monetary policy” when fiscal actions are not. And under today’s conditions, it’s not even clear that the Fed does move last: the Fed is hesitant, and does not appear likely to fully offset the money demand effects of fiscal policy.

  13. Gravatar of Andy Harless Andy Harless
    12. December 2010 at 09:54

    Reserves are a medium of account, as their price is fixed. T-bills are not a media of account.

    This is an arbitrary institutional distinction. As TMDB points out above, one could just as correctly describe the par value of T-bills as the unit of account. T-bills do happen to be sold as discount bonds, but in principle they could just as easily be sold as coupon bonds. And there is a tiny speculative element in their pricing, but what if there weren’t? What if T-bills were non-negotiable coupon bonds? Would they then be a medium of account? And would the Treasury then be conducting monetary policy by issuing T-bills? It seems like too small a distinction on which to hang the definition.

  14. Gravatar of OGT OGT
    12. December 2010 at 10:12

    Scott thanks for taking on this post. As an outsider I enjoy many of the more theoretical discussions because they really help clarify some the issues.

    My inclination is to look at the broader monetary measures that include close but not perfect substitutes for ‘base money,’ because M2 and the now discontinued M3 obviously track much more closely with economic performance than M0. But, it’s in that substitution relationship that things seem to get very interesting, as the degree of substitution is clearly not completely stable.

  15. Gravatar of scott sumner scott sumner
    12. December 2010 at 11:38

    Andy, It is possible that the Fed will shift to an IOR policy instrument, but that doesn’t change anything I said about transmission mechanisms. All it means is that the Fed will be controlling money demand, not money supply. Of course monetary theory is symmetrical, a 10% increase in money supply (holding real demand constant) has an identical impact to a 10% fall in real money demand (holding money supply constant.) In each case there is a level of excess cash balances equal to 10% of the money stock, which will raise the price level by 10%.

    I also don’t agree about the importance of fiscal policy under IOR. The Fed will need some sort of policy goal. Assume it is 2% expected inflation. In that case fiscal policy will have no effect on the price level, as markets will expect the Fed to offset any potential impact in the near future. In this sense the Fed always moves last. It has far more resources than fiscal policymakers, and it moves much more frequently than fiscal policy.

    [I should say that I am ruling out the absurd case where fiscal policymakers issued bonds to buy and hoard currency. That really would be an interesting battle!]

    You said;

    “And under today’s conditions, it’s not even clear that the Fed does move last: the Fed is hesitant, and does not appear likely to fully offset the money demand effects of fiscal policy.”

    I don’t want to be dogmatic here, I’m not arguing fiscal policy has no effect. I’d say that during normal times the fiscal multiplier is near-zero, because the Fed targets inflation. Right now there may be a modest impact on NGDP, as the recent tax cut announcement did boost inflation expectations by about 8 basis points. But even in that case the fiscal action did get partly sterilized by the Fed, as the fed funds futures market responded with a significantly higher probability of a fed funds rate increase in late 2011.

    Regarding T-bills, it is not enough for their price to be fairly stable to be counted as a medium of account, it must be absolutely fixed, and yet a free market price. Lots of products have fairly stable prices. If T-bills did actually have a constant price, then they could be considered a sort of currency. But for that to occur the quantity of each denomination would have to be demand determined, just like any other form of currency. Because T-bill yields are generally different from zero, I think it is better to treat them as money substitutes, which can have important effects on money demand when rates are close to zero (or when the Fed pays IOR.)

    I’m not sure anything of importance hangs on this question. Both our views would allow for a large increase in T-bills to be inflationary. In your view it increases the money supply, in mine it decreases the demand for base money. It just seems cleaner to me to define the medium of account as that asset whose price is always one by convention.

    OGT, Thanks. But I’d add that there are times where M0 tracks economic performance better than M2 or M3. The first year of the Great Depression was one such time, I seem to recall that late 2007 and early 2008 was another.

  16. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    12. December 2010 at 14:46

    Scott, You said:
    “BTW, I am not arguing that changes in the price of T-bills don’t have an important impact on the demand for money-obviously it can, at low interest rates.”
    Demand spillovers between T-bills and monetary base occur both at high and low interest rates.

    The definition of M0 is extremely arbitrary, I think that FDIC insured demand deposits are economically equivalent to the monetary base. T-bills are much more money-like than large denomination time deposits that were included in M3. It is possible to work with any monetary aggregate as long as you track all those demand spillovers.

    You said:
    “I don’t agree about T-bills as the par value is not equal to the market value. To be a medium of account the market price must be fixed. Otherwise all of monetary theory breaks down. The value of money is no longer 1/P.”
    I’m very comfortable with more flexible monetary theory that doesn’t break down when the value of the medium of exchange is equal to the value of the medium of account only at the maturity date, that doesn’t break down when IOR is paid on reserves, that doesn’t break down when cash note numbers participate in the national lottery etc.

  17. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    12. December 2010 at 15:44

    Scott, the current institutional arrangements mean that the base money and the face value of T-bills are equivalent media of account. Those arrangements can break:
    Case 1. Ron Paul abolishes the Fed, but does not succeed in switching to commodity standard. The par value of T-bills will become the sole medium of account.
    Case 2. Ron Paul refuses to lift the debt limit, as a result, some T-bills default. Base money becomes the sole medium of account.

    I’m not satisfied with the current institutional arrangements. I would prefer arrangements where Treasury can default, but the Fed can keep NGDP on target, as only the base money is medium of account.

  18. Gravatar of scott sumner scott sumner
    13. December 2010 at 17:07

    123, Even if DDs are similar to M0 (and they often earn interest, which is an important difference), I still prefer M0, as the Fed controls M0, whereas they only influence M1. I’m not saying using DDs is wrong, just less convenient.

    123, You said;

    “I’m very comfortable with more flexible monetary theory that doesn’t break down when the value of the medium of exchange is equal to the value of the medium of account only at the maturity date, that doesn’t break down when IOR is paid on reserves, that doesn’t break down when cash note numbers participate in the national lottery etc.”

    That’s fine, but we’d have to create entirely new monetary models. You can’t just use existing models of monetary economics, and plug in assets whose nominal value is not equal to one. The output would be nonsense. Obviously anything can be called money. You could call define the dollar as one brick. But it would no longer be true that a doubling of the price level meant the value of money fell in half.

    I still think the base is the most convenient definition, but I can see why some people might prefer currency. I cannot see any pragmatic arguments for any other asset. When people say T-bills are close substitutes for cash at low rates, my response is “fine, let T-bills be one factor that influences the demand for M0.”

    T-bills are not media of account, as their price is not equal to one. The fact that on rare occasions the price may be close to one doesn’t help, the same is true for items sold in “Dollar” stores. Suppose the price of Hershey candy bars had been $1 for the past two years—would you argue they had become a medium of account?

  19. Gravatar of scott sumner scott sumner
    13. December 2010 at 17:09

    123, Why not call the face value of 30 year bonds the “medium of account?”

  20. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    14. December 2010 at 04:28

    Scott, You said:
    “Even if DDs are similar to M0 (and they often earn interest, which is an important difference), I still prefer M0, as the Fed controls M0, whereas they only influence M1. I’m not saying using DDs is wrong, just less convenient.”

    My original point was that M0 theory and M0+DD theory will give you the equivalent results. The same applies to the M0+Tbill theory.

    You said:
    “You could call define the dollar as one brick.”
    Perhaps it is possible to make the M0+bricks theory work, but it would be pointless, as bricks do not facilitate transactions.

    You said:
    ” But it would no longer be true that a doubling of the price level meant the value of money fell in half.”
    It would be an excellent sixth-grade math exercise to determine how much the value of money falls if the the price level doubles. With IOR such exercises are the fact of life, but perhaps this is a good thing, as the MV PQ formula is too simple to be taken seriously by the Keynesians :)

    You said:
    “Why not call the face value of 30 year bonds the “medium of account?”
    Well, yes, today I would say that the face value of all bonds (including 30y) is the medium of account. This is the direct consequence of the fact that the monetary and fiscal policies are closely integrated (too closely for my taste). Where the monetary authority is less integrated, things are different. The face value of Greek 30 year bond is not the medium of account in the Eurozone, although it was four years ago.

  21. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    14. December 2010 at 04:33

    Scott, You said:
    “I still think the base is the most convenient definition, but I can see why some people might prefer currency. I cannot see any pragmatic arguments for any other asset. When people say T-bills are close substitutes for cash at low rates, my response is “fine, let T-bills be one factor that influences the demand for M0.””
    Even though all those theories are equivalent, free-market ideological considerations might push some people towards broad money aggregates. Otherwise money just disappears without the Fed.

    With IOR T-bills are very close substitutes even at the high interest rates.

  22. Gravatar of Lorenzo from Oz Lorenzo from Oz
    15. December 2010 at 17:15

    Bill: thanks. I am glad I asked the question.

  23. Gravatar of Doc Merlin Doc Merlin
    17. December 2010 at 01:01

    @Bill Woolsey:
    You don’t need deposits or credit at all for an electronic money system. Look at BitCoin. It is an electronic money system with neither deposits nor credit nor banks.

  24. Gravatar of scott sumner scott sumner
    18. December 2010 at 05:27

    123, You said;

    “My original point was that M0 theory and M0+DD theory will give you the equivalent results. The same applies to the M0+Tbill theory.”

    This doesn’t respond to what I said. The Fed doesn’t produce DDs. And a huge increase in T-bills has a vastly different impact from a huge increase in the MB. Even if cash and T-bills are currently close substitutes, the public doesn’t expect them to always be close substitutes.

    You said:

    “Perhaps it is possible to make the M0+bricks theory work, but it would be pointless, as bricks do not facilitate transactions.”

    Money is not money because it facilitates transactions. Money is money because it is the medium of account.

    You said;

    “It would be an excellent sixth-grade math exercise to determine how much the value of money falls if the the price level doubles. With IOR such exercises are the fact of life, but perhaps this is a good thing, as the MV PQ formula is too simple to be taken seriously by the Keynesians”

    Sorry, but this comment has nothing to do with my comment that you are reacting to. My comment had nothing to do with MV=PY. Even Keynesians completely agree that the value of money is inversely relate to the price level. It has nothing to do with IOR either.

    You said;

    “Well, yes, today I would say that the face value of all bonds (including 30y) is the medium of account. This is the direct consequence of the fact that the monetary and fiscal policies are closely integrated (too closely for my taste). Where the monetary authority is less integrated, things are different. The face value of Greek 30 year bond is not the medium of account in the Eurozone, although it was four years ago.”

    I can’t tell whether you are joking here. The sine qua non of being a medium of account is having a fixed nominal price. No serious economist would accept T-bonds as a unit of account.

    You said;

    “Even though all those theories are equivalent, free-market ideological considerations might push some people towards broad money aggregates. Otherwise money just disappears without the Fed.”

    Not true, even under a free banking regime there is likely to be a medium of account (such as gold.)

  25. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    19. December 2010 at 07:13

    Scott, You said:
    “The sine qua non of being a medium of account is having a fixed nominal price.”
    The face value of government bonds has a fixed nominal price.

    Let’s say there is a prehistoric community where cattle is money. All the contracts are expressed in terms of five year old cows, but younger cows are a legal tender too, although their price in terms of five year old cow is flexible and market-based. Are younger cows money? Yes, although they do not have a fixed nominal price. Young cows are money because they facilitate transactions, and there is a reliable process that converts them into the standard five year old cows. It is the same with US government bonds.

    You said:
    “Not true, even under a free banking regime there is likely to be a medium of account (such as gold.)”
    What about a free banking regime that is based on fractions of NGDP?

    You said:
    “And a huge increase in T-bills has a vastly different impact from a huge increase in the MB.”
    In both cases the Fed will increase the interest rates.

  26. Gravatar of ssumner ssumner
    19. December 2010 at 12:28

    123, Your last point was proved wrong in late 2008, when they vastly increased the base and interest rates fell from 2% to 1/4%.

    A medium of exchange is not useful if it doesn’t have a fixed nominal price. Not only can you not spend cows at Walmart, you can’t even spend euros at Walmart. If your theory was correct then someone could spend euro currency at Walmart.

    The par value is meaningless, indeed all historical prices are meaningless in economics. What matters in transactions is the asset’s current price. And if the price isn’t fixed, at one, it’s simply not a good medium of exchange or medium of account. That’s why money is special.

  27. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    21. December 2010 at 04:22

    Scott, You said:
    “Your last point was proved wrong in late 2008, when they vastly increased the base and interest rates fell from 2% to 1/4%.”
    Three possible replies:
    1. Ceteris paribus did not hold in late 2008
    2. We know that the increase of the base was way too small
    3. If the monetary base did not change in late 2008, and the equivalent increase was in T-bills, the changes in the interest rates would be the same.

    You said:
    “A medium of exchange is not useful if it doesn’t have a fixed nominal price. Not only can you not spend cows at Walmart, you can’t even spend euros at Walmart. If your theory was correct then someone could spend euro currency at Walmart.”
    Flexible prices of medium of exchange raises transaction costs. But these costs are tiny in large transactions with T-bills. And you can use euro debit cards all over the world without any problems.

    “The par value is meaningless, indeed all historical prices are meaningless in economics. What matters in transactions is the asset’s current price. And if the price isn’t fixed, at one, it’s simply not a good medium of exchange or medium of account. That’s why money is special.”
    Par value is meaningful because it is the promised value of the bond at the maturity date. Where monetary and fiscal authorities are integrated, this promise is backed by the central bank, so the par value of government bonds is one of the units of account.

  28. Gravatar of ssumner ssumner
    22. December 2010 at 19:13

    123, It seems with your first comment you are agree with me that more money doesn’t mean the Fed will raise rates.

    On your second point, if you use a euro debit card at Walmart, you are still paying in US dollars. Then your bank later converts your dollar debt to euros. So that doesn’t change anything I said.

    The big transactions are of trivial importance, most currency is used in small transations. And it is changes in the supply and demand for currency that determeine the price level. If the Fed doubles the currency supply they will probably roughly double the price level. If you double the supply of eurocards nothing happens to the US price level. If you double the number of T-bills nothing happens to the US price level, Instead, the price of T-bills changes.

    I admit that non-medium of exchange assets are used in big transactions, but those assets don’t affect the price level (very much.)

    You said;

    “Where monetary and fiscal authorities are integrated, this promise is backed by the central bank, so the par value of government bonds is one of the units of account.”

    This is impossible, because the unit of account is an abstraction. No object can be a unit of account. If you mean medium of account, you are wrong because by definition the price of a medium of account must be fixed in nominal terms, and that’s not true for bonds. It must always be fixed, not just at maturity. Otherwise changes in the supply of the asset will change its current nominal price–which is impossible for a medium of account.

    In any case it’s a moot point, as it you got your definition accepted, I’d just have to come up with a new term for the object I am talking about, an asset whose nominal market price is always fixed by convention. Call it a numeraire.

  29. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    29. December 2010 at 01:20

    Scott, big transactions are very important in determining the prices of financial assets. The price of financial assets is related to AD via Tobin’s Q.

    You said:
    “If you mean medium of account, you are wrong because by definition the price of a medium of account must be fixed in nominal terms, and that’s not true for bonds.”
    Would your analysis of the Great Depression change if the dollar was fixed to gold with a 2% trading band? No. So the theories with more flexible definition of medium of account are useful. For example, these days in the Eurozone Germand bonds are the new gold, even though they have a fixed nominal price only on the maturity date.

    You said:
    “If you double the number of T-bills nothing happens to the US price level, Instead, the price of T-bills changes.”
    This will happen because the Fed will reduce the growth of monetary base. How is that different from the scenario where the supply of $100 bills doubles and the supply of $50 bills is reduced instead?

  30. Gravatar of Scott Sumner Scott Sumner
    30. December 2010 at 09:11

    123, How do transactions affect asset prices? Do big tranactions make asset prices go up or down? I had thought expectations drove asset prices. But I do agree that asset prices are important for AD.

    I still don’t agree on the medium of account, except in one special case. If the yield on T-bills is near zero, and is expected to stay near zero forever, then I agree with you. In that case a $10,000 T-bill is indistinguishable from a 100 $100 bills. But if T-bill yields are expected to rise to 4% in future years, that means they won’t be influencing the price level in future years. Since current prices are closely linked to future expected prices, that means T-bills also will have little current effect on AD, unless they impact the demand for currency.

    In my final example I was assuming the amount of base money is held constant, and the supply of T-bills doubles. I still say it has little effect on the price level (unless yields are driven to near zero.)

  31. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    1. January 2011 at 05:16

    Scott, in the old days repo was a component of M3, even though repo was used only in the big financial transactions. It all makes sense, as the same excess cash balances mechanism is still operating. If the supply of very liquid T-bills doubles, investor portfolios have too high proportion of liquid assets. Prices of other financial assets increase until investors are indifferent between holding new T-bill securities and selling them for other financial assets.

    Regarding the medium of account, in those countries where fiscal and monetary authorities are integrated, and we know most central banks except ECB fall into this category, reserves and T-bills are in a very tight equilibrium, and expected risk and economic return are identical.

    You said:
    “In my final example I was assuming the amount of base money is held constant, and the supply of T-bills doubles. I still say it has little effect on the price level (unless yields are driven to near zero.)”
    Fiscal multiplier is zero when monetary policy is active. But in your example, it is passive. It doesn’t matter if yields are high or low, but fiscal policy has strong effect on AD in your example if we exclude the cases where sovereign default risk is high.

  32. Gravatar of Scott Sumner Scott Sumner
    1. January 2011 at 14:06

    123, I don’t recall assuming monetary policy is passive, but if I did then I agree that fiscal plicy can have a modest effect on AD. However I think monetary policy is usually active.

    I guess if you integrate monetary and fiscal policy things get much more complicated. But in the US those two policies are not integrated, so increases in T-bills are not inflationary here.

  33. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    2. January 2011 at 01:41

    Scott, You said “In my final example I was assuming the amount of base money is held constant”. I think this is equivalent to an assumption that monetary policy is passive.

    In the US integration of monetary and fiscal policy today is quite useful. As fiscal authorities keep the default risk of fiscal liabilities very low, the expected risk and expected return on T-bills and reserves are identical (this is true now, and this was true before they started paying IOR). Compare this to the situation in Greece, where the expected risk and return on ECB liabilities is very different from the expected risk and return on Greek T-bills.

    Increases in T-bills are not inflationary as the Fed adjusts money supply accordingly.

  34. Gravatar of ssumner ssumner
    4. January 2011 at 10:32

    123, I don’t see why it’s useful for reserves and T-bills to have the same expected return.

  35. Gravatar of 123 – TheMoneyDemand Blog 123 - TheMoneyDemand Blog
    5. January 2011 at 05:30

    Scott, usually this just means that both the central bank and the treasury are expected to remain solvent.

    Now there are various systems that ensure that reserves and T-bills have the same expected return. Corridor system is better than the system where reserves have no explicit yield, but their quantity is artificially constrained so their holders earn a convenience yield.

  36. Gravatar of interfluidity » Belated write-up of AEA/AFA meeting, Part I interfluidity » Belated write-up of AEA/AFA meeting, Part I
    22. January 2011 at 22:59

    […] Andy Harless (but see Sumner’s rejoinder), I think the distinction between fiscal and monetary policy has grown very blurry. Monetary […]

  37. Gravatar of TheMoneyIllusion » Waldman on monetary and fiscal policy TheMoneyIllusion » Waldman on monetary and fiscal policy
    24. January 2011 at 12:55

    […] me.  I was reminded of this when reading a new post by Steve Waldman: Like Andy Harless (but see Sumner’s rejoinder), I think the distinction between fiscal and monetary policy has grown very blurry. Monetary […]

  38. Gravatar of Andy Harlessの金融政策論 by Scott Sumner  – 道草 Andy Harlessの金融政策論 by Scott Sumner  – 道草
    29. January 2011 at 19:19

    […] Harlessの金融政策論 by Scott Sumner  //  Scott Sumnerのブログから”Andy Harless on monetary policy”ï¼ˆ11. December 2010)の翻訳紹介。 ä»¥å‰ç´¹ä»‹ã—たAndy […]

  39. Gravatar of Skepticlawyer » Put not your faith in labels: naming in the dismal science – guest post by Lorenzo Skepticlawyer » Put not your faith in labels: naming in the dismal science – guest post by Lorenzo
    11. September 2011 at 08:44

    […] it massively reduces transaction costs. Prices and obligations can be dealt with by a single medium of account across all transactions. We have a single unit of account to record prices and obligations in and a […]

  40. Gravatar of Skepticlawyer » Easy Guide to Monetary Policy Skepticlawyer » Easy Guide to Monetary Policy
    6. June 2012 at 21:48

    […] of account. Something that is used as a medium of exchange and embodies the unit of account is a medium of account and so money; something used to both quantify and pay exchange […]

  41. Gravatar of Skepticlawyer » The real convenience of money Skepticlawyer » The real convenience of money
    28. March 2013 at 21:42

    […] next step was to create media of account; things that were a medium of exchange that also instantiated the unit of account. That is […]

  42. Gravatar of Skepticlawyer » Why gold/currency ratios mattered in the interwar period Skepticlawyer » Why gold/currency ratios mattered in the interwar period
    5. December 2015 at 13:45

    […] gold standard sets a value in gold to the unit of account, making gold the medium of account. So, no matter how many (for instance) currency dollars are in circulation, the value of gold sets […]

Leave a Reply