Let the ECB capital key float

Bankers clear and settle with each other at a clearing house

Perry Mehrling had in interesting comment about how to settle the Eurosystem's Target2 imbalance problem.

If there were Eurobills, balances could be settled periodically by transfer of assets, just as is done in the Federal Reserve System. More precisely, if there were a System Open Market Account at the ECB, in which all of the national central banks held shares, settlement could be made by transfer of shares.

Perry is talking about adapting the structure of the Fed's Interdistrict Settlement Account to Europe. To understand the ISA, check out my Idiot's Guide to the Federal Reserve Interdistrict Settlement Account. In short, the 12 regional Reserve banks run up debts and credits to each other over the course of the year due to changes in payments flows. These debts and credits are settled each year by transferring securities that have been bought in open market operations from debtor Reserves banks to creditor Reserve banks.

The Eurosystem, on the other hand, doesn't require that the clearing debts wracked up by the various National Central Banks (NCBs) be settled. Which is odd. Not even Keynes's ICU would have allowed infinite debts, and Keynes was a forgiving sort of guy.

Perry's idea is that with the Eurosystem embarking on a program of Outright Monetary Transactions (OMTs), maybe the securities amassed will allow for a mechanism to settle intra-Eurosystem imbalances. Debtor NCBs like the Bank of Greece will have to transfer OMT securities to creditor banks like the Bundesbank. Perry specifically discusses the idea of having each NCB own shares in the OMT portfolio and have these shares transferable so as to facilitate settlement.

I like this idea and think it can be twinned with the already existing ECB capital key. What is the capital key?

All NCBs have an ownership stake in the overlying ECB. The relative amounts held by each are determined by the capital key. The key's weights are based on relative population and GDP. Germany for instance, has been given an 18.9% weighting in the capital key. Greece has been given a 1.96% weighting.

The ECB holds a unique set of assets on its balance sheet (see year-end 2011 statements). These have been transferred to it from the NCBs. First, it holds 16 million ounces of gold, worth around €21b. It also holds around €41 in forex reserves denominated in yen, dollars, and more. Lastly, it holds a few assets purchased during previous open market campaigns. Also worth considering is that the Eurosystem's total profits are paid out according to the capital key. This means that the profits yielded by assets held by the Bank of Greece don't necessarily get paid out to the BoG... they get amalgamated with all NCB profits and then allocated to each individual NCB according to the capital key. So having a big weighting in the capital key is definitely worth something.

Say that going forward all OMT purchases are conducted through the ECB and not the NCBs. That means that in addition to its gold and forex position, the ECB gets even bigger, and shares in the capital key are rendered more meaningful.

You can then institute a Fed-style settlement program by letting each NCB's weighting in the capital key float. Each year debtors get their share in the capital key lowered. Creditors get their's increased. That adds some quid pro quo to intra-Eurosystem balances. What happens if a central bank's representation in the capital key were to fall below zero because of persistent capital flight? Then the national government would have to recapitalize their respective NCB's contribution to the ECB so that it's portion of the key rises back to 0%. Or you could soft-pedal the whole thing and institute some sort of broad system of reforms a country has to initiate once it falls below 0%. Either way, the system is redesigned to have a tendency to equilibrate.

Is such a tendency necessary? I leave you with some words from Keynes's Proposals for an International Currency (or clearing) Union, February 11, 1942:

Measures would be necessary to prevent the piling up of credit and debit balances without limit, and the system would have failed in the long run if it did not possess sufficient capacity for self-equilibrium to prevent this.

The free banking alternative to the centralized provision of lender of last resort services

Inspired by Perry Mehrling and Fischer Black:

I think I'd take your future ideal financial model even further (slide 9). The C5 in your model provides what you call liquidity puts. I see no reason why these liquidity puts need be provided by a central bank. In the future, financial products called liquidity options - the option to buy or sell some asset (say Apple stock) at a guaranteed point in the bid-ask spread - would be popular financial products traded on organized exchanges. Just as Apple CDS allow investors to split off Apple credit risk and distribute it across the economy, so would Apple liquidity options split away the liquidity risk of transacting in Apple stock in the secondary market and evenly distribute this risk to those willing and capable of holding it.

A private liquidity options market has some advantages over a monopoly last resort system. Liquidity would be competitively priced and no longer supplied in an opaque manner. Central banks would either vacate the market for liquidity services or price their liquidity products off the private liquidity options market. Subsidies to or penalties on institutions anxious for liquidity insurance would be a thing of the past.

If central banks were to cease providing liquidity options, their sole role in the future would be as managers of the clearing and settlement system. The provision of paper money can be easily fulfilled by private banks. I guess central banks would also have to manage the price level.

The above is a free-banking view of the world. The lender-of-last resort role is transferred from central banks to private markets. It is distilled into just another financial product.

Thinking in terms of stocks: From Fisher to Fischer

In an older post, Scott Sumner had an interesting comment:

The most recent inflation rate in Greece is 1.7%, whereas Spain has 1.9% inflation. I don’t know about you, but I find those figures to be astounding. That’s not deflation, and yet Tyler’s clearly right that they are being buffeted by powerful deflationary forces. I’d make several observations:

1. This shows the poverty of our language. Economics lacks a term for falling NGDP, even though falling NGDP is arguably the single most important concept in all of macro, indeed the cause of the Great Depression. So we call it “deflation” which is actually an entirely different concept. I wouldn’t be the first to find connections between the poverty of our language and the poverty of our thinking.

Fisher Black's dream

Perry Merhling had an interesting quote from Fischer Black:

Thus a long term corporate bond could actually be sold to three separate persons. One would supply the money for the bond; one would bear the interest rate risk; and one would bear the risk of default. The last two would not have to put up any capital for the bonds, although they might have to post some sort of collateral

In the comments I pointed out that a fourth person can be added to the list - someone who bears the liquidity risk of that corporate bond in the secondary market. This would amount to a liquidity option. Essentially, the bearer of liquidity risk would allow the bond owner to sell that bond at some preset level in the bid-ask spread. For instance, the option could allow the bond owner to immediately sell their entire bond holdings at the upper end of the spread, or at the ask price. Normally, sellers can only sell quickly if they accept a price near the bid price, or lower end of the bid-ask spread. When illiquidity strikes and spreads widen, usually because buyers depart and there are less bids, an option to sell at the ask price rather than the bid price increases in value.

IMF and SDRs

A few thoughts on the IMF and SDRs over at the Money View.

The IMF and the SDR program are difficult to de-consolidate:

As far as I know, the IMF doesn't run the SDR program on its own balance sheet, it just administers the SDR program. Using your example, the EU and the US issue promises to currency which are held in some mutual account managed by the IMF, and that account in turn issues SDRs back to the EU and US. So in effect, the IMF doesn't swap its own promises with the EU and the US. Rather, the US and EU are swapping promises with each other with the IMF as facilitator.

That being said,

The last time I checked, the IMF was the largest owner of SDRs, all held on its own account. 

Here is the current distribution of SDRs across nations and institutions.

Target2 and the Federal Reserve Interdistrict Settlement Account

Perry Mehrling wrote an interesting post called Why did the ECB LTROs help?

He visits the comparison of the Federal Reserve Interdistrict Settlement Account and the ECB Target2 settlement mechanism. I have also found this comparison in a number of other publications, see the list at bottom. Here is the comment I left at the Money View blog.

My reading of the Euro-system rules is that deficit national central banks (NCBs) never have to settle with surplus NCBs. These intra-system balances can grow ad infinitum. Thus, deficit NCBs don't have to worry about owning acceptable assets for settlement, since there is no ultimate day of reckoning. The result is that survival constraints for Eurosystem NCBs are far looser than the survival constraints faced by regional Federal Reserve banks, which must settle each year. In the old days this settlement was conducted by transfers of gold certificates amongst regional Federal Reserve banks. After 1975 the settlement medium was switched to securities held in SOMA. But in the case of the Eurosystem, there is no ultimate settlement mechanism or medium that I am aware of, neither gold nor securities. 

Merhling's co-blogger Daniel Neilson had an interesting reply in which he pointed out that it would appear that the Fed's April settlement seems to no longer be occurring:

Indeed the Federal Reserve System seems to have decided to let FRBNY accumulate a large claim on the other Reserve Banks. One can see why the various Reserve Banks might not feel the need to imagine a breakup of the United States in deciding on the makeup of their balance sheets.

The appearance of non-settlement is noticeable in Mehrling's chart. The reversion to mean that should be occurring each April settlement did not happen in 2011. Debts to the New York Fed from other regional Federal Reserve banks are, seemingly, being allowed to accumulate.

Whether in fact settlement rules are being ignored is a mystery to me. Here is my comment:

Odd that this is not mentioned in the Fed's financial notes. What is the resolution of the data in that chart? End of week? It might not give sufficient granularity for us to assess whether claims were allowed to accumulate, or were settled for a day or two before returning to trend.

Buiter explains why the constraint can be escaped:

"The Interdistrict Settlement Account must be settled once a year with gold-backed securities or Federal treasury bills. This would represent a constraint on inter-district credit flows only if the stock of Federal Treasury bills allocated to the individual Federal Reserve banks was exogenous. However, individual regional reserve banks can always buy Federal treasury bills from banks or other holders of the stuff in their own districts, financing this with an increase in base money. The Federal Reserve Board could then decide to undo this transaction or ‘sterilise’, it. However, an interest-rate-setting Fed will only undo this, if the regional Fed’s Treasury bill purchase and the associated increase in base money were to lead to an excessive divergence between the actual Federal Funds rate and the Federal Funds target. This is highly unlikely. Even when the Fed’s official policy rate is at the effective lower bound for the official policy rate and the Fed is engaged in QE, there is no effective constraint on the ability of regional Reserve Banks to settle the Interdistrict Settlement Account imbalances with Treasury Bills funded with base money issuance. The yearly settlement requirement in the Interdistrict Settlement Account procedure would thus not appear to be a constraint on persistent credit imbalances between individual Federal Reserve banks’ districts."

I am hoping to investigate this mystery a bit further.

There are a number of conflicting resources explaining how the interdistrict account is settled, and how binding this is.

1. The key source is pg 136 - 138 of the Financial Account Manual for the Federal Reserve.
2. The Buiter quote I mention above comes from a Citi report called Original Sinn.
3. Peter Garber's Deutsche Bank article The Mechanics of Intra Euro Capital Flight also touches on the comparison.
4. Citi prints Hans-Werner Sinn reply to Buiter here, giving his views on interdistrict settlement.
5. Here is Sinn's initial paper.
6. I got the reference to the 1975 settlement changes from Meltzer, A History of the Fed, Vol 1. "The diminished stock of gold and gold certificates and rising levels of reserves and deposits required a change in interbank settlement... In 1975 the operations staff recommended that monthly gold transfers cease. Reserve banks other than New York would change once a year. New York would pay for withdrawals and receive deposits from the Treasury. Once a year, the Interdistrict Settlement Fund would reallocate securities in the System Open Market Account to balance accounts. Gold remain as collateral for the note issue, but securities would be the principal collateral (memo, Maurice McWhirter and Alan Holmes to FOMC, Board Records, April 11, 1975). Step by step, gold lost its monetary role and main provisions of the 1913 Federal Reserve Act disappeared."

There are no other good sources that I've been able to locate on the Target2/interdistrict comparison.

Note: I have blogged more extensively on the Interdistrict account here.

The Borges Problem part II

Nick Rowe writes a post called Macroeconomics and the Celestial Emporium of Benevolent Knowledge.

It is a conjunction of a bunch of themes he has touched on in the last few weeks, including the great debt debate and categorization.

I asked him:

In your previous post, you advocated adopting the most "useful" way of dividing up the world into categories. But what does that mean? Useful in terms of teaching students, reaching the layman, articulating theory amongst other economists, calculating statistics, conducting policy?

Are you saying that economists should have multiple "categorical universes" in their head? Or is their "one ring to rule them all" way of dividing up the universe that you are trying to tease out?

One useful reason for having one standardized way of splitting the world into categories is it makes conversation easier. Having multiple ways adds subtlety but confusion.

I am learning that the core of economics (or at least near to it) is about words. We create meanings for things in the real world using some of our existing words, and this creates categories, but since other words can also be used to create similiar-though-not-overlapping categories, we have to choose between categories. This process comes prior to mathematical econ, since the math needs categories like C and I before mathematical phrases can be put together. 

The last bit about mathematics reminds me of an exchange I had with Daniel Kuehn a while back about the use of words and math in economics. 

Here is Nick's response:

I'm not sure on the answers to all your questions. Even if we all agreed on what the true theories were, we might want one theory for one purpose and a different theory for another purpose. And each theory would have its own most useful set of categories.

Yep, it might make conversation harder. But maybe some conversations would be impossible if we had to stick to one set of categories that didn't fit the topic at hand.

I can't disagree with him. Is this an argument in favor of heterodox economic? After all, entire schools can form around a divergent way of categorizing the world. Should economists be fluent in multiple economic viewpoints? The blogosphere is surely a great place for that. Actually, this reminds me of a recent Perry Mehrling post on which I participated. Here is Mehrling:

...it is important to emphasize that each of these heterodox schools exist, and persist, because it is organized around some essentially correct insight about how the banking system works... The problem is that, so far as I can see, each of the heterodox schools has part of the truth, not the whole thing. The same could be said about the orthodoxy against which the heterodox schools define themselves... We don't therefore want to choose which school to belong to; rather we want to determine which of these correct insights provides the most useful explanatory frame for whatever issue is currently at hand. Today it might be one; tomorrow it might be another. That is what the debate is about, or should be anyway.

See an earlier and related post.

ECB, NCBs, collateral, capital key, Target2, and intra-eurosystem credit

Two comments on The Money View. One on Perry Mehrling's The IMF and the Collateral Crunch and the other on Daniel H. Neilson's Is there an ECB?

Neilson links to the erroneous Tornell/Westerman piece. My comments on this are in a previous post. In short, Karl Whelan's Worse than Sinn clarifies the issue. Sterilization by the Bundesbank is not happening. 

Merhling and me discuss the nature of the transactions conducted between borrowing NCBs and the lending ECB.

Perry, I can't find any explicit reference to whether intra-Eurosystem credits are collateralized or not.

But I still think not. Collateral is posted by a borrower to a lender to protect the lender should the borrower default. Then the lender can collect the collateral instead.

But ECB losses are dealt with in a specific way. See bottom of http://www.ecb.int/ecb/orga/capital/html/index.en.html

In short, if the ECB suffers a loss on a loan to an NCB then that loss is allocated to all NCBs according to the ECB's capital key.

The March 2011 Bundesbank report describes this:

"An actual loss will be incurred only if and when a Eurosystem counterparty defaults and the collateral it posted does not realise the full value of the collateralised refinancing operations despite the risk control measures applied by the Eurosystem. Any actual loss would always be borne by the Eurosystem as a whole, regardless of which national bank records it. The cost of such a loss would be shared among the national banks in line with the capital key."
So it would be redundant or unecessary for an NCB to post collateral to the ECB for ECB credit, since in the case of non-payment the ECB has a claim on all NCBs to make up the loss.

That being said, I don't think this necessarily changes the thrust of your post.