« Alpha Banks, Beta Banks, and negative rates | Main | University budget surpluses: irreversible investment and uncertain demand »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Nick,

I’m somewhat puzzled by your construction of the two policies under which silicon and paper standards are considered for their effectiveness.

According to your choices, the policy that has the more aggressive easing policy objective in terms of a silicon money inflation target (3 per cent instead of 2 per cent) is the policy that has the less aggressive easing action in terms of silicon interest rates and corresponding paper depreciation rates (presumably 0 per cent instead of (1) per cent).

That seems counterintuitive and even contradictory as a construction for alternative policy choices.

Maybe I'm missing something, but I find it difficult to try and figure out which policy is better for either silicon or paper standards because of this.

"Plus under the second policy there would be the extra hassle of firms and their customers having to translate from posted paper prices into silicon prices if they paid using silicon rather than paper."

JKH, how would firms and their customers pay with silicon (central bank reserves)?

I agree with JKH. In the context of avoiding the lower bound, inflation is a benefit and you would want silicon to depreciate even with menu costs. Adding currency depreciation is to prevent silicon from turning into currency. There is some slack before encountering this as currency currently subsidies silicon to the extent of merchant fees and rebates so already carries a penalty (~3% in the old days but hard to believe it is that much now). It would take greater inflation than that before there would be much demand for currency.

JKH: I'm assuming a standard (simple) New Keynesian macro model. That means I'm assuming the central bank hits its inflation target (at least on average) under both policies (both policies are equally "effective" in that sense), and that real interest rates (on silicon money) will be the same under both policies. Since the inflation rate (measured against silicon money) will be 1% higher in policy 1 than in policy 2, the nominal interest rate on silicon money will therefore also be 1% higher in policy 1 than in policy 2.

Nick,

I understand your last point, I think. This is tangential, but it does seem strange to me that there exists a choice between two monetary policies with two different nominal rates of interest (on silicon) that can end up at the same assumed real rate of interest – i.e. no effect or at least no differential effect of policy on the equilibrium real rate, when one of the policies seems to include more active easing from an interest rate perspective.

Apart from that, it seems like there are a lot of moving pieces to be considered in mapping these monetary policy choices to medium of account choices. But given a policy with negative rates and paper depreciation, and given the use of either medium of exchange, an exchange rate that is not 1:1 ends up being messy in any case.

JKH: the assumption that different target inflation rates have exactly zero effect on real interest rates is unlikely to be exactly true. (Especially since the nominal rate of interest on paper currency is pinned down at 0%.) We can think of various reasons why different inflation rates have various real effects (like the menu costs I mention in this post), which is why the choice of inflation target matters. And in general equilibrium, almost everything depends on almost everything else, so exact superneutrality for real interest rates is implausible. But it's an outcome of simple models, and is often used as an approximate working assumption, and not wildly at odds with the long run data.

But there's an important difference between *getting to* a higher inflation rate and *being at* a higher inflation rate. If expected inflation is slow to adjust, you may need to lower nominal interest rates initially to get to higher inflation, but then raise them above where they were initially once you have got there. That's why nominal interest rates are such a confusing measure of monetary ease or tightness.

"Apart from that, it seems like there are a lot of moving pieces to be considered in mapping these monetary policy choices to medium of account choices. But given a policy with negative rates and paper depreciation, and given the use of either medium of exchange, an exchange rate that is not 1:1 ends up being messy in any case."

Yes. That is presumably why using foreign money (or two different monies) is relatively rare.

The reason that paper and silicon can both be considered 'money' today is that consumers can treat them as equivalents.

This equivalence would be broken by your 'crawling peg'. Businesses are willing to work with the fiction that paper and silicon are equivalents because their transaction costs are relatively low, on the order of a couple of percent. Moreover, that transaction wedge is bounded – when some cards like American Express are too expensive for businesses to pretend equality, they simply stop accepting them.

Your crawling peg would make it progressively more difficult for businesses to feign equality, and history suggests that when equality breaks it's cash that wins out. Going back to the early days of retail silicon, we see that personal checks (proto-silicon) were either accepted at par or not accepted at all, never (almost never?) accepted at a discount.

A crawling silicon peg might work if we again break apart alpha and beta silicon, such that beta silicon retains par with alpha paper and beta silicon depreciates with respect to alpha silicon, but now we're discussing prices in terms of a money (alpha silicon) that can never be used for exchange: businesses would deal with alpha paper or beta silicon only.

We can imagine that for historical reasons prices are marked in terms of a MoA that is no longer issued (Guinea/Pound), but it strains reason to imagine that this is the case for a MoA that doesn't have an effectively fixed exchange rate with respect to the MoE.

I am struggling to compare the two alternative policies. The two policies are more like pictures from two different worlds.

"inflation" is a direct-exchange measurement taken at discrete time periods. In other words, using any two named items of value, the exchange rate is found to be variable when measured using identical time periods. Therefore we see "inflation" to be a comparison of values.

Next, take a look at the currency-reserve world. In this world, we accept that new money is created when a bank loan is made and spent. In this world, money is destroyed when loans are retired by payoff.

If loans create money, then how are the two types of money used in the currency-reserve world created? By loan creation but what would be the ratio of creation between the two types? Under your proposed scenario, the value ratio would change so we would be introducing a managed inflation between the two types of money. That should change the value of the underlying loans as well.

Concluding, what policy might be best to "mitigate the risk of hitting the Zero Lower Bound"? Well, while I see pictures from two worlds, neither seems to be related to interest rates and the zero lower bound. It seems to me that interest rates are the cost of money charged by the lender. The lender makes that decision and if the central bank wishes to lend at zero rate of interest--it can.

Majro: you might be right. If you are right, that means that my second policy is off the table. The Central Bank either raises the inflation target, or else takes steps that (whether intended or not) lead to the elimination of paper.

Depreciating paper is analogous to foreign currency in this case.

There's no fundamental reason why retailers wouldn't accept both silicon and paper.

I can pay retailers in Mexico quite easily in either pesos or dollars. And their FX rate changes every day.

The MOA is pesos, with both fixed and non-fixed MOE accepted.

@JKH:

> The MOA is pesos, with both fixed and non-fixed MOE accepted.

That's fine and not unexpected. But that's not analogous to this situation, where "alpha silicon" is not a medium of exchange. Are there any stores in Mexico that price in Pesos but accept only US Dollars?

Majromax

Beta silicon is the silicon medium of exchange for retail transactions.

Beta silicon is fixed to alpha silicon.

Dollar beta silicon and peso beta silicon are fixed to dollar alpha silicon and peso alpha silicon respectively.

From the previous discussion, banks are the ones maintaining the beta silicon pegs. They can choose to exchange it 1:1 for alpha paper or alpha silicon.

Beta silicon and alpha paper serve as the media of exchange for retail transactions. Thus far, retail outlets use alpha paper as their medium of account even with ostensible parity, as evidenced by the lack of separate pricing even as institutions moved from taking only alpha paper to both alpha paper and beta silicon. (Canadian case in point, Tim Hortons didn't often take credit/debit a decade or so ago; now that they do there's neither a cash discount nor a debit/credit fee.)

Also look at it from a minimization-of-effort point of view: if alpha paper remains the MoA and banks peg beta silicon to alpha paper, all of the "currency conversion" is centralized in those relatively few banks. If beta silicon is pegged to alpha silicon and floats against alpha paper, then every individual and business will have to perform these conversions.

JKH said: "Beta silicon is fixed to alpha silicon."

How does that work?

Beta silica can be exchanged at a fixed rate of 1 to 1 for alpha currency paper at a commercial bank.

Alpha silica can be exchanged at a fixed rate of 1 to 1 for alpha currency paper at the central bank but only by commercial banks.

Is there a *direct* market to exchange beta silica and alpha silica?

Majromax said: "From the previous discussion, banks are the ones maintaining the beta silicon pegs."

Is there any reason it is not both ways?

Roger: "I am struggling to compare the two alternative policies. The two policies are more like pictures from two different worlds."

Think of them as two parallel worlds, with two different policies, but all the exogenous things the same (so both worlds get hit by the same shocks etc.).

The (unstated) assumption is that paper money must pay the owner 0% nominal interest, for practical reasons. But silicon money can pay the owner any (positive or negative) rate of interest the bank issuing it chooses (but people will stop holding silicon and will hold paper instead if that rate gets much below 0%). The rate of interest on loans (that people get paid for holding non-money IOUs) is a third rate of interest.

Ok, I think I see. This avoids the lower bound by letting rates go negative which is easy enough on silicon until payments begin bouncing. With lower inflation, they will spend more time negative, though I would say it is less about a choice of a negative rate on currency in normal times than when even the higher inflation rate would need it, fully expecting they will not be able to meet their objectives otherwise.

Banks would fix beta silicon to alpha silicon – because alpha silicon is the ultimate settlement medium for the banking system and because the interest rate on alpha silicon (either the interest rate on reserves, or the interest rate for interbank lending of reserves) is the basis for all other interest rates in the banking system. All rates are developed as spreads and yield curves from this rate as the origin rate. That’s why the market is so interested in central bank interest rates in the first place. Conversely, banks have no interest in alpha paper other than to carry an inventory of it to serve the interests of non-bank customers. The dollar volume of banking activity which settles ultimately in alpha silicon (including cheques, credit card, and automated transfer transactions) is a vast multiple of the dollar volume of alpha paper that non-banks and individuals use to settle retail transactions. There is no way that banks would fix their beta silicon to depreciating alpha paper.

Similarly, most retailers would have no problem posting primary prices in alpha/beta silicon terms. It is only the small convenience operators who would have a real issue, and even they have a non-trivial and likely growing credit card payment business. It would be easy enough to post beta silicon prices along with a percentage premium available for calculation of corresponding alpha paper rates. Registers can be programmed to do it with a daily input premium price and a touch of the button. That piece is analogous to retailers accepting payment in either domestic or foreign currency. They do it all the time in Mexico using hand held calculators.

The history of cash as the conventional medium of account is simply not relevant when things reach the stage of accelerating digital volume and particularly in an environment of negative interest rates and a managed peg for depreciating alpha paper.

Lord: yep.

What JKH is saying makes sense to me. Though he is more confident than I am that the silicon standard will prevail. But I think JKH would agree that the only right answer is "it depends", and that in other countries, or in Canada past, where the banking system is less developed and more people use alpha paper, the answer might be "paper standard".

One other thing, and I think this may not be inconsistent with what Nick just wrote above, although I'm not sure he'll agree with it:

In a normal developed system without negative interest rates and without paper depreciation, there is really no issue of differentiating between beta silicon and alpha paper as the designated medium of account. They are structurally the same value. Exceptions are relatively minor.

That means that denoting alpha paper as the medium of account in normal circumstances is essentially arbitrary and/or biased – it reflects a historic continuity but it is really not meaningful in present day terms.

So in the case of advanced economies at least, with reasonably sophisticated banking systems, there really is no fundamental reason not to claim that beta silicon is just as much the medium of account as alpha paper - and perhaps even more so.


Further to my second last comment, merchants could charge for alpha paper transactions in the form of a premium over beta silicon pricing that would look similar in form to an HST (tax) charge.

TMF September 01, 2016 at 02:14 PM

“Is there a *direct* market to exchange beta silica and alpha silica?”

arguably yes

Commercial banks bring claims on beta silicon (for example cheques) to the interbank clearings.

A cheque on beta silicon is essentially a moving form of beta silicon - its certainly not alpha silicon or alpha paper

Those cheques are exchanged for credits and debits in alpha silicon.

E.g. Royal brings a cheque drawn on BMO to the interbank clearings.

That’s BMO silicon in effect.

Royal gets an alpha silicon credit (increase in its reserve account).

BMO gets an alpha silicon debit (decrease in its reserve account).

That’s a marketplace for exchange of alpha and beta silicon, in effect.

It’s direct in terms of the flow of beta silicon in exchange for a flow of alpha silicon.

Although Royal and BMO need to ensure internally consistent stock/flow accounting for beta silicon.

i.e. in conjunction with the flow of cheques in exchange for alpha silicon accounting entries, Royal credits a beta silicon customer account and BMO debits a beta silicon customer account

JKH: I think I agree with that. We simply can't tell the difference, at least in normal times. But in historical cases where a bank suspends convertibility, so its monetary liabilities trade at a discount to the central bank's monetary liabilities, then we can tell the difference.

JKH said: "In a normal developed system without negative interest rates and without paper depreciation, there is really no issue of differentiating between beta silicon and alpha paper as the designated medium of account. They are structurally the same value. Exceptions are relatively minor.

That means that denoting alpha paper as the medium of account in normal circumstances is essentially arbitrary and/or biased – it reflects a historic continuity but it is really not meaningful in present day terms.

So in the case of advanced economies at least, with reasonably sophisticated banking systems, there really is no fundamental reason not to claim that beta silicon is just as much the medium of account as alpha paper - and perhaps even more so."

I think I agree with that too.

Next, assume all commercial banks are solvent.

Are there any other conditions when a central bank will not act as a lender of last resort to the commercial banks?

JKH said: "A cheque on beta silicon is essentially a moving form of beta silicon"

I would say a check is instructions on what do with demand deposits. That means to move them.

"- its certainly not alpha silicon or alpha paper"

I agree with that.

"Royal gets an alpha silicon credit (increase in its reserve account).

BMO gets an alpha silicon debit (decrease in its reserve account)."

"Although Royal and BMO need to ensure internally consistent stock/flow accounting for beta silicon.

i.e. in conjunction with the flow of cheques in exchange for alpha silicon accounting entries, Royal credits a beta silicon customer account and BMO debits a beta silicon customer account"

I agree with those.

"Those cheques are exchanged for credits and debits in alpha silicon."

"That’s a marketplace for exchange of alpha and beta silicon, in effect.

It’s direct in terms of the flow of beta silicon in exchange for a flow of alpha silicon."

I am not sure I would characterize it like that.

Assume I have a BMO checking account and a Royal checking account. For whatever reason, I want to move demand deposits from my BMO checking account to my Royal checking account. I could go to BMO and get currency to physically move it. I take it to Royal and deposit it. I'd rather not do that. I write a check and present it to Royal. That check means please move my demand deposits from my BMO checking account to my Royal checking account for me. Someone needs to physically move vault cash from BMO to Royal. The banks would rather not do that either. Instead the fed moves central bank reserves (vault cash substitutes) for them electronically.

In both cases, there is a markdown of demand deposits both as an asset and liability at BMO. In both cases, there is a markup of demand deposits as an asset and liability at Royal. In both cases, there is some kind of movement of either vault cash or central bank reserves.

I don't really see a *direct* exchange of demand deposits (beta silica) and central bank reserves (alpha silica).

How does that sound?

We have in effect a self imposed corralito.
https://en.wikipedia.org/wiki/Corralito
Since we now longer use cash, we can't have the same problems as Argentina had in the early '00s.
But then there are 4 kinds of countries...
https://mostlyeconomics.wordpress.com/2014/02/20/there-are-four-kinds-of-countries-in-the-world-developed-countries-undeveloped-countries-japan-and-argentina/

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad