One of my regular commenters J. P. Koning recently drew my attention to this post by Nick Rowe about the “identity” (to use a somewhat loaded term) between the gold standard and the CPI standard (aka inflation targeting). Nick poses the following question:
Is there any fundamental theoretical difference between how monetary policy worked under the gold standard and how monetary policy works today for a modern inflation-targeting central bank?
I will just note parenthetically that being a Canadian and therefore likely having been spared from listening to seemingly endless soundbites of Newt Gingrich pontificating about “fundamental this” and “fundamental that,” and “fundamentally this” and “fundamentally that,” Nick obviously has not developed the sort of allergic reaction to the mere sound or appearance of that now hopelessly hackneyed word with which many of his neighbors to the south are now incurably afflicted.
To attack the question, Nick starts with the most extreme version of the gold standard in which central bank notes are nothing but receipts for an equivalent amount of gold (given the official and unchangeable legal conversion rate between bank notes and gold. He then proceeds to relax the assumptions underlying the extreme version of the gold standard from which he starts, allowing central bank reserves to be less than its liabilities, then allowing reserves to be zero, but still maintaining a fixed conversion rate between central bank notes and gold. (Over 25 years ago, Fischer Black developed a theoretical model of a gold standard with (near) zero reserves. See his paper “The Gold Standard with Double Feedback and Near Zero Reserves” published as chapter 5 of his book Business Cycles and Equilibrium. Whether anyone else has explored the idea of a gold standard with zero reserves I don’t know. But I don’t dispute that a gold standard could function with zero reserves, but I think there may be doubt about the robustness of such a gold standard to various possible shocks.) From here Nick further relaxes the underlying assumptions to allow a continuous adjustment of the legal conversion rate between central bank notes and gold at say a 2% annual rate. Then he allows the actual conversion rate to fluctuate within a range of 1% above to 1% below the official rate. And then he allows the base to be changed while keeping any changes in the base unbiased so the expectation is always that the conversion rate will continue to rise at a 2% annual rate. Having reached this point, Nick starts to relax the assumption that gold is the sole standard, first adding silver to get a symmetallic standard, and then many goods and services to get a broad based standard from which a little addition and subtraction and appropriate weighting bring us to the CPI standard.
Having gone through this lengthy step-by-step transformation, Nick seems to think that he has shown an identity between the gold standard and a modern inflation-targeting central bank. To which my response is: not so fast, Nick.
What Nick seems to be missing is that a central bank under a gold standard is operating passively unless it changes its stock of gold reserves, and even if it does change its stock of gold reserves, the central bank is still effectively passive unless, by changing its holdings of reserves, it can alter the real value of gold. On the other hand, I don’t see how one could characterize an inflation-targeting central bank as acting passively unless there was a direct market mechanism by which the public forced the central bank to achieve its inflation target. If a central bank did not maintain the legal conversion rate between its bank notes and gold, it would be violating a precise legal obligation to engage in a specific set of transactions. Instead of buying and selling gold at $20.67 an ounce, it would be buying and selling at some other price. If an inflation-targeting central bank does not meet its inflation target, can anyone specify the specific transactions that it was obligated to make that it refused to make when called upon to do so by a member of the public? And this is aside from the fact that no one even knows whether an inflation targeting central bank is achieving or not achieving its target at the time that it is conducting whatever transactions it is conducting in pursuit of whatever goal it is pursuing.
In short, an inflation-targeting central bank cannot be said to be operating under the same or an analogous set of constraints as a central bank operating under a gold standard, at least not under any gold standard that I would recognize as such.
Uneasy Money 
David:
Unaccustomed as I am to agreeing with Nick, he is right that an inflation-targeting central bank issues its dollars just as passively as does a gold standard bank. Start with a bank that has issued 100 paper dollars, against which it holds 20 oz of gold as reserves, plus bonds worth 80 oz. The bank targets $1=1 oz. If the street value of $1 falls to .99 oz, the public will bring in dollars to exchange for EITHER 1 oz of gold or 1 oz. worth of bonds. The bank could refuse to exchange, but then it would be abandoning the target, which is contrary to the initial assumption. It seems clear that in this case, a bank that targets $1=1 oz, and always exchanges dollars for either gold or bonds at that rate, is a passive issuer of dollars.
Now, instead of 20 oz. of gold, let the bank hold 20 CPI baskets of goods, plus bonds worth 80 CPI baskets. The bank targets $1=1 basket. If the street value of $1 falls to .99 baskets, then customers will bring in dollars to exchange for either baskets or 1 basket worth of bonds. Just as above, a bank that targets $1=1 basket, and always exchanges dollars for baskets or bonds at that rate, is a passive issuer of dollars.
Nick would probably go on to say that the bank could issue $100 without holding a full 100 oz worth of assets against it. He says something like 30 oz. would be enough, but he might be wavering.
LikeLike
Mike Sproul: surely a gold standard is based on an existing stock of gold. Clearly the bank can acquire more gold, or change its backing ratio, but, either way, it is an existing stock of gold that the standard is based on. So, if gold production grows faster than general output, you get inflation and if it grows slower, you get deflation (adding in normal caveats about shifts between monetary and non-monetary uses of gold).
Inflation targeting is not based on an existing stock of anything. It is based on expectations about output and the money supply. The CPI “basket” is not remotely like the stock of gold, it is just a representative benchmark. Shifts in output affect how the bank operates in a way it does not with a gold standard. Hence an inflation-targeting central bank can hit its inflation target over time, even though it may wander around it in any given time period. (And, if it is the RBA and has an “average over the business cycle target”, it can smooth out output fluctuations because it does not have to follow contractions downwards. Indeed, as far as I can see, an “average over the business cycle target” is, effectively, an NGDP target.)
Thus the point that Thomas Sargent made when he said that the euro was like a gold standard for participating countries. It was so because they had no control over the value of their currency or its supply. This is a very different situation than even the most rigid inflation-targeting national central bank.
LikeLike
Well, I have thought more than a bit about Black-type systems.
And further, about free banking evolution where gold reserves fall to zero.
I think you are wrong about central banks being “passive” with a gold standard.
There task is always to adjust current monetary conditions (the current excess supply or demand conditions for money) so that the gold market clears at “par.”
They are trying to avoid losing reserves (and should be trying to avoid gaining reserves.)
It is true, that when they make errors, and when they hold reserves, the change in the demand for reserves by the central bank (or private banks in a free banking system) directly impacts the supply and demand conditions in the gold market and so allows the gold market to clear, but assuming they were holding the desired reserves before, they have to reverse that change. And avoiding that is the point.
With a gold standard with zero reserves, or a standard using a broad bundle were reserves are impractical and so zero, then this effect doesn’t work and instead all that you have is the monetary authority (or private banks in free banking versions) adjusting current monetary conditions so that the market clears.
LikeLike
David: off the top of my head: with 100% gold reserves, central banks could be 100% “passive”. The legislated mechanism “allow the public to buy or sell money for gold at the fixed price” is *all* they have to follow. But as soon as we move away from 100% gold reserves, central banks must actively choose their tactics if they are to avoid running out of reserves. So they cannot be 100% passive.
If inflation targeting central banks held 100% reserves of CPI baskets, they could be 100% passive too. It’s fractional reserves of the targeted good that means central banking requires some degree of active choice and skill.
On re-thinking my post, I think the biggest difference is that gold prices are very flexible while the CPI is composed of many sticky prices. So there are longer lags with CPI targeting than gold price targeting.
(Sorry about “fundamentally”. I did stay one night in Newt’s home town a couple of years back. Lovely place. French (Canadian?) origins, IIRC.)
LikeLike
BTW, I just did a new post, partly based on what you said about Wicksteed. It’s ages since I read Wicksteed, and I hope I haven’t misrepresented his view too much.
LikeLike
About Nick’s steps from this to that to this etc. I think that the most important benefit of competitive free banking is that it would allow things to evolve more quickly than our Government run monetary system can evolve. I do not know how it would have ended up but I bet had we had a free banking system all along the problem of gold and a contraction due to a run on would have been solved. If a bank says that its currency will no longer be redeemable in gold people can use a different bank’s currency but they might realize that the bank’s loans are valuable assets and so continue to use its currency.
LikeLike
Nick/David, I think you need to agree on the definition of “passive”. I’m going to guess that what David means is that if a CB is to be qualified as passive, then the stock of money it provides is determined by the quantity demanded by the public at the rate set by the CB.
Incidentally, I had so much fun in the comments section of Nick’s and David’s posts that I posted a summary of my take here:
http://jpkoning.blogspot.ca/2012/04/you-say-hot-potato-he-says-endogenous.html
LikeLike
I would see inflation more contained under an Gold Standard. Therefore, the central bank should have an easier job at controlling inflation.
LikeLike
JP: “I’m going to guess that what David means is that if a CB is to be qualified as passive, then the stock of money it provides is determined by the quantity demanded by the public at the rate set by the CB.”
If by “rate” in that sentence you mean “inflation rate”, that would be true (in the medium term, when inflation is on target) under inflation targeting. The equilibrium stock of money will be determined by: the target inflation rate chosen by the central bank; a load of other things determined by the public that we could summarise as “the demand for money”.
LikeLike
I see some billionaires are planning mine gold from asteroids. Also, gold can be produced from mercury through some sort of nuclear bombardment.
What will the gold nuts do when gold becomes plentiful?
Will we all be rich?
LikeLike
Nick, that sounds right to me. In the gold standard sense, “rate” would mean the price at which the CB will convert liabilities into gold. In that case the equilibrium stock of money will be determined by the public’s demand for CB money at the $/oz rate set by the CB.
LikeLike
“If an inflation-targeting central bank does not meet its inflation target, can anyone specify the specific transactions that it was obligated to make that it refused to make when called upon to do so by a member of the public?”
It seems to me that the transaction it failed to make was to purchase (sell) the appropriate amount of financial assets from (to) the open market. ie. if CPI rises above its target, the CB should have been selling more financial assets.
If I had 100% faith in a central bank’s promise to hit its targets, were CPI to come in at 6% while the target was just 3%, I’d immediately sell assets and hold cash in anticipation of future central bank purchases of cash. I’d be able to front-run the CB and earn easy profits. I’d have the most faith in, say, the rather extreme Reserve Bank of New Zealand which ensconces policy targets in its governing act. The Governor can be fired for failing to meet targets.
LikeLike
Mike, In principle, a central bank could base its monetary policy on a legal commitment to make its currency convertible into commodity basket. If a central bank adopted such a regime, there would be arbitaguers out there ensuring that the value of the basket remained equal to the stated target or at least within the range of fluctuation allowed by transactions costs. In that case, I would agree that the behavior of the central bank would be passive. I don’t regard that case as being at all similar to the Fed targeting the CPI, because there is no set of arbitrage transactions that becomes profitable when the Fed fails to meet its CPI target and simultaneously force the central bank to meet its CPI target or go into default as a result.
Bill, My point is that as long as there is a gold standard, the bank is legally obligated to provide a fixed weight of gold if it is presented a banknote. It can do so either by holding reserves of gold or by purchasing gold as needed to redeem banknotes. Failure to do so would mean that the bank was in default, so either the bank needs to hold reserves or it needs to calibrate its policy so that it is not called upon to redeem its notes. A CPI targeting central bank is under no such constraint.
Nick, My point is that convertibility gives the public a means of enforcing the central bank’s commitment. It may choose its tactics, how much reserves to hold, but it cannot allow the conversion price to change and cannot refuse to honor its commitment without being in default. There is no comparable mechanism by which an inflation target becomes self-enforcing.
Floccina, A private competitive system does allow more scope for experimentation and unplanned evolutionary development. However, as I argued in my book Free Banking and Monetary Reform, competition does not lead to the adoption of the optimal monetary standard. That choice needs to be made at a higher level to achieve social optimality.
JP, Well the rate set by the CB is a matter of its own discretion, so I don’t know if that is the key difference. I am now repeating myself, but to say it one more time, under a gold standard the central bank is forced by the decisions of the public to follow a policy that is consistent with its obligation to convert its banknotes into gold. Under inflation targeting, the public has no analogous means of forcing the central bank to comply with its inflation target.
Tas, Under a gold standard, a central bank cannot control the rate of inflation except by controlling the real value of gold. The rate of inflation depends on whether gold is becoming more or less valuable in terms of all other commodities.
Benjamin, This is just science fantasy.
JP, That is my point. Under the gold standard, the appropriate transactions are imposed on the central bank by the decisions of the public to demand redemption or cash in gold for banknotes. Under inflation-targeting, the public has no recourse by which to engage in transactions that force the central bank to pursue a policy consistent with inflation targeting.
LikeLike
David: “There is no comparable mechanism by which an inflation target becomes self-enforcing.”
Under present mechanisms for targeting inflation, I agree. Because what the central bank is targeting is its own internal forecast of the CPI, and that internal forecast isn’t really observable. Perhaps if the central bank directly targeted a futures contract on the CPI it might be different?
LikeLike
David: On reflection, your point is well-taken. But does it really matter “fundamentally” (sorry) for monetary theory? At the level of monetary policy tactics, maybe it does. At the level of monetary policy strategy (i.e. assuming the central bank has the right tactics to hit its strategic objective) I don’t think it does.
LikeLike
Nick (4/26/2012 @ 7:40 pm), Totally! In chapter 10 of my book Free Banking and Monetary Reform, I proposed a system (devised by Earl Thompson) of indirect convertibility to stabilize the expectation of a wage index via a futures contract issued by the Fed. An expectation of a rising wage index would trigger purchases of the contract and decreasing the quantity of currency and non-interest bearing reserves thereby reducing expectation of the wage level while an expectation of a falling wage index would trigger sales of the contract with the opposite effect on the expectation. Scott Sumner came up with the same idea independently, and the system would work automatically. But there has to be a contractual commitment that the central bank is undertaking in order to get the system to work automatically. The gold standard provides such an enforceable contractual commitment, though it has undesirable properties under certain circumstances that make it a bad choice for a monetary standard.
Nick (4/27/2012 @ 4:43 am) Not sure what to make of this follow-up. I don’t think I get your point. Which is disappointing after I thought after your previous comment that we were finally on the same page.
LikeLike
David: we are on the same page. I understand your @2:18 and agree. You must have misunderstood my @4:43 in some way. No biggie.
LikeLike