Gold Standard Technical Operating Discussions 3: Discretion Vs. Automaticity


January 29, 2012


Now, at last, we come to the topic which prompted this rather interesting discussion of technical operating mechanisms.

January 15, 2012: Gold Standard Technical Operating Discussions 2: More Variations
January 8, 2012: Some Gold Standand Technical Operating Discussions

For some reason, I seem to become inspired at the beginning of the year. I'll start on some innocuous topic, and it turns out to be something quite important.

January 30, 2011: Italy With the Gold Standard 1861-1914
January 23, 2011: The Gold Standard in Britain 1778-1844
January 9, 2011: The "Money Supply" With a Gold Standard 2: 1880-1970
January 2, 2011: The "Money Supply" With a Gold Standard

March 23, 2008: How Banks Work 7: the Lender of Last Resort
March 16, 2008: How Banks Work 6: Liquidy Crises and Bank Runs
March 9, 2008: How Banks Work 5: Selling Loans
February 24, 2008: How Banks Work 4: Banks and the Economy
February 17, 2008: How Banks Work 3: More Elephant Poop
February 10, 2008: How Banks Work 2: Shitting Like an Elephant
February 3, 2008: How Banks Work

The topic is the role of discretion vs. automaticity in the daily operations of a gold standard system.

We began our discussions with a typical currency board arrangement. (Actually, I think even currency boards don't hold very much base money, but probably use a demand deposit or maybe short-term debt of a foreign government.) A typical currency board is wholly automatic as regards to adjustments of base money supply. Every action is prompted by private market participants ("PMPs") wishing to buy or sell with the currency board authority.

However, the currency board does have some discretion regarding the composition of its reserve holdings. We examined how that could work.

You can set up a gold standard system in this way too. Our first two gold standard examples were of this type. Their actions to reduce or increase the base money supply were prompted entirely by PMPs wishing to trade with the gold standard authority.

Historically, there were no doubt systems of this type in use somewhere, during the past two centuries. However, the core systems -- Britain, the U.S., Germany and so forth -- were generally of the hybrid type. This was the last type we looked at, in which the gold standard authority has an automatic-type mechanism in the form of bullion redeemability, and also a discretionary mechanism in the form of open market operations in high-quality debt, typically domestic government bonds. We saw that you could also develop a system that uses open-market operations in bonds entirely, and does not have a redeemability element, but set up a system of automatic rules of operation. I don't know of any example of this historically, but it certainly could be done.

The hybrid systems in use historically typically used open market operations, on a discretionary basis, as the first and preferred avenue of adjusting the monetary base. The practical reasons for this are obvious enough: they tended to hold most of their reserves in the form of debt, instead of bullion, because this would maximize seinorage income. A ratio of around 80% debt:20% bullion was common. Second, it is quite a lot easier to buy and sell debt than to transport bullion, which has higher transaction costs.

Thus, a typical system would have bullion "buy/sell" points, whether official or the natural result of the transaction costs of bullion, let's say around 2% on either side of the parity ratio. In other words, if the official parity was $1000:one troy oz, then people would go to the central bank to buy bullion (redeem banknotes) at a market price of $1020/oz., and sell bullion (trade for banknotes) around $980, perhaps. In other words, when the market price was $1020/oz., you could bring $1000 to the central bank and get an ounce of gold in return, thus generating a $20 profit, minus transaction costs.

However, the gold standard authority ("GSA") would often act to adjust the monetary base, through unsterilized purchases and sales of bonds, before the value of the currency reached these "bullion points." If the market price was $1005, in other words, it took a little more than $1000 to buy and ounce of gold and therefore the value of the currency was a little low, the GSA would sell some of its bond holdings, thus reducing the monetary base and supporting the value of the currency. It would drift back toward its $1000/oz. parity, ideally never reaching the $1020 bullion redemption point. Or, if the market didn't move back toward its parity, the GSA would sell more bonds, reducing the monetary base further. If that still didn't work, then the unsterilized redemption into bullion would act as yet another means to reduce the monetary base.

The timing and size of these open market operations in bonds were left to the discretion of the GSA and its operators. Over time, they probably developed a natural feel for appropriate timing and size. That's the idea, anyway.

However, this presents some complications. It relies upon a certain level of mastery of the GSA operators. Unfortunately, as we have seen, sometimes these people have no idea what they are doing. They might do exactly the wrong thing, buying when they should be selling, or selling when they should be buying! Instead of improving upon an automatic system of redeemability, now we are instead undermining it. This happened in the late 1960s in the U.S.

Also, any open market operation, in bonds or bullion, changes the monetary base and thus would probably change the reserves of banks. (The exception would be if banknotes were redeemed.) This would have some effect on the market for overnight bank loans, which is actually not that big a deal -- there are many ways of funding other than overnight loans. We have way too much fixation on this today. However, the discretion to make open market operations leads naturally to the discretion to manage short-term interest rates. In the 1920s, the Fed began to do this. As it was managing the gold standard system, it could also, concurrently, subtly manage the short-term bank loan market. This was nothing like what we have today, but you can see how the seeds were planted.

I'm a bit of a traditionalist. There are a few good reasons for this, but in general I think there is way too high of a priority placed on "originality" or "innovation" today. I use quotes because often these ideas are not very original or innovative. There are a lot of reasons for this excessive focus on novelty. It is, for example, a Heroic Materialist theme. We are ardent believers in "newer and better." We just assume that the iPhone 4 is better than the flip-phones of five years ago, and we emphasize the same kind of "innovation" in all spheres, whether it is appropriate or not. This then relates to the academic world, where careers are built upon the impression of intellectual leadership. Ambitious professor types feel that they can more easily obtain tenure by inventing some newfangled notion, even if total nonsense as it often is, rather than just saying: the old way worked well, so why not just use that? I'm much more practical, and would rather go with a way that works than have to sift through the dozens of bonehead propositions that people come up with these days.

Thus, when asked how to design a gold standard system for today, I tend toward the kind of hybrid system that worked in the U.S. and Britain for several successful decades and even centuries.

However, I admit that this has certain drawbacks, as I have outlined above. This might be a time for some improvements, namely a wholly automatic system in which there is no discretion.

Of course, there is always discretion in how the automatic system is set up. Also, there will always be a "manual override," either de facto or de jure. But, maybe the day-to-day operations of the system should not be left to anyone's supposedly good judgement, because we all know how rare that is.

So, let's think of some wholly automatic systems. We've already outlined two. One is the bullion-redeemability-only system. There are no open market operations in debt, except in the role of the "lender of last resort" which is really a sort of overlay system. The monetary base is adjusted entirely through unsterilized sales and purchases of gold bullion, initiated by PMPs wishing to transact with the GSA.

Within this system, the GSA then has the option of deciding how much of its reserve to hold in either gold or gold-linked bonds. It can adjust this as necessary, through transactions that do not change the monetary base, as we saw in our previous items in this series.

We also looked earlier at various ways to set up a wholly automatic system that used only open-market operations, and did not have bullion redeemability. For example, you could say that if the market price is 1% away from the parity price, then the GSA would increase or decrease the monetary base by 0.5% via open market operations, per day. If the currency is 1% or more below its parity value for four days, the GSA would reduce the monetary base by selling bonds by 0.5% each day, for a total of 2.0%. Over a month, about 23 working days, you would reduce the monetary base by 11.5%, which is quite a lot actually.

You could add a laddered sequence of activity. If the parity price deviated by 2% or more, then the GSA would adjust the monetary base by an additional 1.0%, per day, for a total of 1.5%. This would add up to 7.5% over the course of a week, which is a lot.

There we have two wholly automatic systems. Now, we could develop some hybrid systems, that have both automatic elements in them.

In a hybrid system, the basic question is how the two elements would work together. Would the first avenue of action be an open market bond operation? Or gold redeemability? Or would we like them to operate simultaneously?

Let's look at examples of all three.

In our first example, we will have open market operations in bonds as our first means of operation. This corresponds to the U.S./British example, but with an automatic rather than discretionary system. So, let's say we have our 0.5% base money adjustment triggered by a 1% or more deviation from parity, per day, and then gold redeemability (or monetization) at a 2.0% deviation from parity.

In our second example, gold redeemability is the first means of operation. We could make a rule that gold redeemability shall be the normal course of operations, but if gold reserves fall below 10%, or if gold transactions (either redemptions or monetizations) amount to more than 2% of total base money in a week, then additional open market operations in bonds will be added at a rate of 1% per day.

In our third example, we could have both operate simultaneously. For example, for every $1 million of gold that is either redeemed or monetized ("monetized" means gold is sold to the GSA at the parity price), we will match with another $1 million of open-market bond operations. Or, you could do it in proportion. If gold reserves are 20% of total reserves, then each gold transaction will be matched with an automatically-triggered open market bond transaction of four times the size. This would maintain the reserve ratio at a stable 20%.

You could develop a number of other variations as well.

All in all, I tend to think that some sort of wholly-automatic hybrid system involving both gold redeemability and open market bond operations would be best. However, people need to learn how to do this before they can design, establish, and maintain such a system. It's not really that hard, but we are starting from a very rudimentary level of understanding today.