« Interest rates, exchange rates, and the Bank of Canada | Main | What is the right sort of risky asset for central banks to own? »


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

The greatest threat to a central bank is a run on its gold reserves. By being known to hold safe and liquid assets such a run is less likely And if it does occur, these assets can be sold off relatively quickly. Since, of course, a central bank is small relative to the entire gold standard world, any liquidity effect of selling off its assets is small. Unless, of course, it is holding risky and illiquid securities to start with!

Obviously, Rowe, you understand nothing about banking. That is why bankers should be put in charge of the lead bank in a country! Economists!

Bill: my morning laugh! You might well be right. It's a holdover from the gold standard, and fixed exchange rates, and accepting the dominant discourse of bankers. But the Bank of Canada is run by economists, not bankers.

The debt that occurs between two currencies is the result of productive differences between the two economies.

If one economy, year-after-year, is willing to ship production to the second currency without balancing trade, a paper debt must accumulate. Some entity must hold this debt.

Now consider WHO might hold this debt.

It could be the maker of BMWs if buying German cars was the cause of Greek debt.

It could be a bank if the bank elected to finance BMW's Greek sales.

It could be the Central Bank if, in the long run, the Central Bank decided to become a local bank in disguise.

The lesson here is that trade between currencies must be balanced if debts between currencies are to be avoided.

"So risky assets are priced at a discount to safe assets"

Only the component directly correlated to the discount rate (systemic risk). The market does not place premia/discount on idiosyncratic risk. Regardless of the std dev or risk tolerance of the marketplace.

Bill Woolsey has it right, I think. The rationales behind a lot of central bank practice are very much post-hoc explanations of practices based on tradition.

In liquid markets there is much less need to do fundamental investment analysis. You just pay the market price. You don't have to be "good" at investing. This is an institutional competence argument for sticking to liquid assets.

As for safety, one can make a fiscal smoothing argument. All else equal you don't want choppy seigniorage revenues, it complicates tax and debt management.

Roger: "The debt that occurs between two currencies is the result of productive differences between the two economies."

No it isn't. If one country (Canada) has lots of investment opportunities and few savers, and another country (the UK) is the opposite, then the first country will borrow from the second.

But the reason why borrowing and lending exist does not affect the topic of this post, which is about who should hold the riskier debt vs who should hold the safer debt.

Derivs: yep, that's the standard CAPM. Though the simple version of CAPM doesn't always work, if asset holders themselves have idiosyncratic risk, and want to insure against that risk, and maybe the central bank faces idiosyncratic risk. But regardless, why should central banks not hold "risky" assets, under any definition you like?

WP: I too think that Bill's explanation is the best, at least so far. (Maybe someone will come up with something better?)

MR: Greetings! If central banks simply held an index mutual fund, they don't need to do fundamental analysis. I think the fiscal smoothing argument is just the same as the consumption smoothing argument for why I myself don't like risk. But why should central banks, as an arm of government, be much more risk averse than me?

Yes but index fund shares are very liquid. We wouldn't want central banks doing private equity investing or making direct illiquid loans. Too much risk of ovepaying.

I'm not sure but I don't think the fiscal smoothing argument is analogous to personal consumption smoothing. I'm thinking about barro / tax smoothing / deadweight costs of taxation. By the way I don't think my argument on this latter point is particularly strong.

I think that this can be explained through reversible policy instruments.

Central banks change the size of the monetary base through open market operations. To increase the monetary base, they buy stuff. To reduce the monetary base, they sell that stuff back to the market.

The problem is that this is only fully symmetric if that stuff that the central bank purchased maintains a steady nominal value. If the central bank bought hog bellies to expand the monetary base, then if the hog market crashed then it would have to invent something new to reduce the monetary base -- there wouldn't be any asset of note to sell.

There are also additional impacts for the fiscal authority's budget constraint. The central bank provides a small but steady amount of seigniorage revenue to the government, but that's based on the difference between returns on its assets and liabilities. If its assets are volatile, then that seigniorage revenue is very unsteady. If it ever goes negative, whereby the CB needs to ask the fiscal authority for a transfer to reduce the monetary base, then that would be a political disaster.

Also, if we consider seigniorage revenue to be a deadweight loss, then we want policy instruments that minimize that revenue. If the central bank purchases risky assets, then on average it will necessarily earn excess return from the risk premium.

We also can think about the use of interest rates as a policy instrument. If we believe anything from a Phillips Curve to the Taylor Rule, then changes in the risk-free nominal interest rate change the experienced risk-free real interest rate, which in turn affect private aggregate demand. If the CB wishes to use this policy lever, then it makes sense to deal primarily in assets that directly influence said lever: purchasing hog bellies has only a very indirect effect on government bond yields.


* Reversible OMOs require assets that maintain known values
* Fiscal constraints make steady seigniorage revenues superior to volatile ones, with asymmetric risks for positive/negative.
* Interest rate arguments suggest that the CB should act on assets most directly tied to the risk-free rate

If the CB buys riskless assets it affects only the risk-free rate of interest and will affect the economy in a neutral way.
If it buys risky assets it also affect risk-adjusted rate-of-interests and this may affect the way that risk is taken into account when investment decisions are made and have undesirable side effects in the real economy.

The risks associated with buying risky assets may be worth it when the effects of buying riskfree assets stop working (becasue their purchase stops affecting interest rates much, which are already very low).

I can,t do better than Bill. Glopoutniks ( unfortunately called Central banks because tradtion and people not understanding they are not banks)) are nwe devices, less thsan two centuries old. They were not invented for a goal according to a theoretical construct (unlike spoutniks). They were a typical british muddle through (the 1825 crisis). Nobody knew they would be needed later. Nobody kbew they would have to stop being a "bank". Nobody thought they were inventing a new thing that needed a new non-confusing name. Sometimes whorfism rule: the word define the thought. People say Central bank but think bank, see Banque Nationale de Suisse.
Those who were the first to need and invent one, the Bank of England and its offshoot in the former Empire , being the first, are now the most advanced in their thinking.
It will be a long and hard slog. Like the Reformation-Counter Reformation that convulsed Europe for three centuries. Like the similar convulsion of another large religion now. In the end, people will understand but a lot of damage will be done by the Landsknehts.
Coffee has not yet kicked in so sorry for the rambling of a humble IO guy.

If the central bank buys a risky financial asset in enough size to move the market, it benefits both the issuer of those assets (who may now be able to raise additional funds at lower cost) and current holders of those assets. If we assume the purchase has no effect on the real economy, then by extension others must have lost out, if for no other reason than that how wealthy you are depends not simply on how much money you have, but how much everyone else has relative to you. (Note that this is not necessarily the equity owner of the central bank, or the people who pay tax to the equity owner.)

So it's a distributional issue and, in general, the more risky the asset, the more potential for benefiting one group at the expense of others. This is not an argument against the central bank holding risky assets at all, but the more it ventures into that area, the harder it becomes to avoid favouring certain groups over others.

MR: I was thinking of the Bank of Canada just buying the S&P/TSX stock market index. It could keep a few more liquid assets for day to day trading. Just like a pension plan might do the same.

Individuals consumption smooth because of diminishing marginal utility of consumption. Governments Barro-Pigou tax smooth because of increasing marginal deadweight costs of taxation. Yes, it's not strictly the same thing, but formally it looks sort of the same.

Majro: if central banks never needed to reverse (buy back their money) they wouldn't need assets at all. Just vary the rate of printing, and put the vacuum cleaner away.

We all face costs if our assets crash in value. Why should the central bank be *more* risk-averse and face bigger costs? Plus, the central bank can always borrow against its future revenue stream if it wants (aka pay interest on reserves).

"If it ever goes negative, whereby the CB needs to ask the fiscal authority for a transfer to reduce the monetary base, then that would be a political disaster."

It could borrow/pay interest on reserves, or just have "negative equity" for a while. It would only need to ask the fiscal authority (its owner) for funds if it had a very big loss and wanted to buy back a very big amount of cash.

Whether it buys government bonds or hog bellies, the money base increases the same amount. A portfolio switch from government bonds to hog bellies may have micro effects, but it's not obvious how it would affect the inflation rate.

Why buy an index fund? The most indexed fund possible is the whole GNP. That's why glopoutnik (sorry CB) should buy the least particular asset, gunmint bonds when things are good. So that gunmint can buy things to give to people (don't call it fiscal policy it would frighten the horses) or cut taxes (I am old, Nick, but the american custom of immediate tax rebate is swell). And we do it through gunmint because people would not understand a "bank" giving money because we have still not changed the letterhead to "Glopoutnik of Canada" even less to Glopoutnik Nationale de Suisse (there is seemingly no word in german for that concept. Maybe no possible word...)

Central banks hold low risk assets for the same reason that old-time note-issuing bankers did: Because their customers like to hold low-risk currency, and low-risk currency must be backed by low-risk assets.


> A portfolio switch from government bonds to hog bellies may have micro effects, but it's not obvious how it would affect the inflation rate.

But isn't that just the point?

If we're thinking in the "most macro" sense, then all we see are the monetary base, aggregate demand, and inflation. From a thousand miles up, all assets look alike, and it really doesn't matter if the central bank buys bonds or hog bellies.

But the central bank doesn't just operate on the thousand-mile-macro level; its officers and the government have particular interest in microeconomic factors. From that perspective, stable seigniorage revenue and least-distorting OMOs have advantages over more volatile alternatives.

> if central banks never needed to reverse (buy back their money) they wouldn't need assets at all. Just vary the rate of printing, and put the vacuum cleaner away.

That might beg the question. Have central banks not needed to buy back their money in part because they have the power to do so, either actively through OMOs or passively through keeping the cash from maturing assets?

Borrowing against future income can't reduce the long-term level of the monetary base (relative to the current level; it can definitely do so relative to a positive trend) if the central bank follows a no-Ponzi condition.

You have previously made the point that central banks with a great deal of credibility don't actually have to take extreme actions, but part of that credibility is through the Chuck Norris effect of being able to take extreme actions if necessary.

As far a non-gov't assets are concerned, the big difference is that shareholders get a vote, while bond holder's don't. CB printing and buying shares in companies would probably not sit well with many for political reasons.

Does the CB have comparative advantage in pricing unusual or distressed assets? No. Quite the opposite, the CB should stick to buying assets with a clear market price lest private interests steal some of those profits for themselves.

"Do central banks have a comparative disadvantage at holding illiquid assets? That just sounds weird, because central banks themselves produce the most liquid assets of all."

Indeed. It's the wrong question, isn't it? It is not "all about comparative advantage" with CBs, is it?

"You can make a case, as I did once, that central banks should sometimes actually prefer holding risky assets, even if other things were equal. But let's set that case aside. And you can make a case, in a liquidity crisis, that central banks should actually prefer holding illiquid assets, even if other things were equal. But let's set that case aside too."

If you eliminate these and similar considerations, what is the point of having a central bank?

But, to play the game, let's assume that CBs would prefer not to buy risky assets. When would it make sense for them to buy them? One possible answer might be that doing so would reduce the risk of those assets, making them a bargain. But that gets into questions that seem to be verboten.

Majromax: "if we consider seigniorage revenue to be a deadweight loss"

Why should we do that? Unlike taxation, it does not destroy options. Au contraire, it creates them.

Because the ECB already owns most of them? Because their perception is markets under price most of these risks, or that they over price them due to unfamiliarity?

Mike Sproul's answer is of course perfectly consistent, within Mike's theory, but that theory is wrong, because central banks give their profits away to their favourite charity, so those forgone profits are what truly "back" the value of the currency.

Baksun. Or rather, back later.


> Why should we do that? Unlike taxation, [seigniorage revenue] does not destroy options. Au contraire, it creates them.

I don't necessarily think that way myself, but nil seigniorage revenue is implied by the Friedman rule. Besides, seigniorage revenue does amount to a tax on currency-holders, relative to the counterfactual of interest-on-money.

So if we are to follow theories that argue for minimal seigniorage revenue, that suggests that the central bank should not be seeking excess returns from risk tolerance.

Political controversy aversion? If a central bank, owned by a government, buys governments bonds, then it doesn't have to justify its portfolio decisions to nosy politicians.

Because they are afraid of confusing the cause with the effect of their own actions, circularity?

"within Mike's theory, but that theory is wrong, because central banks give their profits away to their favourite charity"

It’s probably easier to imagine the government issuing paper money directly to government employees, as wages, with no central bank. Maybe that’s profitable, (if the market rate of interest exceeds the printing/handling cost of the paper) but that wouldn’t disprove my assertion that it’s the government’s assets that back the money. Bringing the central bank into the picture doesn’t change anything, except a new set of books needs to be kept.

"So risky assets are priced at a discount to safe assets, and so yield a higher expected rate of return than safe assets"

You remain incorrect!! Risk does not get priced into the price of an asset. (it's asset pricing 101).

In very rough layman terms, Up and down risk are symmetrical and therefore cancel, leaving expected value of all idiosyncratic assets to be the Rf rate, regardless of std dev or risk tolerance.

This is universally accepted as a fact.


A Chicago boy gets it right. John Cochrane - see bottom of pg.17

Could it be just so that central banks can use interest rate signals more directly?

If I am a central bank and, in order to reach long run inflation targets, I want to immediately reach a target of 0.75% short term rate on safe assets, isn't it just easier to to transactions on short term safe assets at around that rate?

How would it work? Commercial banks would have to buy stocks and sell them to central banks while agreeing to purchase them again at a usually slightly higher price later? If the price of stocks went down significantly, the banks would have an immediate loss and liquidity issues. Wouldn't the banking system as a whole be more vulnerable to stock market swings?

Central banks are an extension of commercial banks and the same question could be asked of commercial banks. Why should commercial banks give out loans instead of buying stocks in their customer's endeavours. Loans are easier to price and value. Imagine if I bought a car and instead of getting a loan I would promise to give 5% of my salary, whatever that may be, for the next five years to the car company?

Could it not just be about pushing the risks as far as possible toward the end borrower or investor who is likely to have more information and do better allocation?

"And remember, the biggest asset that a central bank owns does not even appear on its balance sheet: that's the expected present value of its future seigniorage profits from printing money. "

How does your thinking change with regards to those central banks that endure seigniorage losses. For instance, the central bank of Costa Rica has consistently lost money for the last two decades; its expected PV of future profits is negative.

Because the purchase of risky and illiquid assets has been successfully outsourced to the treasury. Thus, monetary policy is (almost) always right while fiscal policy is often irresponsible.

In line with market fiscalist, odie and others above - wouldn't taking default risk on the CB balance sheet land more in fiscal policy territory than pure monetary policy? Moreover, absorbing the idiosyncratic risk of any one market will distort price and so likely distort market micro outcomes. No good.
Of course in a crisis a CB historically does more than just manage monetary aggregates. Buying illiquid assets or lending to systemically important but illiquid institutions is something CBs have done out of necessity. Think Bernanke's QE 1 purchases of MBS or even his swap lines to other CBs.

Risk management focus and efficiency in a framework of separate institutional responsibilities - to minimize the fiscal effects of any given monetary policy.

Derivs: the standard CAPM defines "risk" of an asset as covariance of the total returns on that asset with the total returns on the market portfolio. But that is not quite right. The more sophisticated version defines "risk" as negative covariance of the total returns on that asset with the marginal utility of consumption.

Do all finance guys teach their grandmothers how to suck eggs?

Max + JKH provides a plausible theory. Central banks break that rule in a liquidity crisis, when they act as lenders of last resort. They could also agree with politicians that central banks would just buy the index. There may be more than one way to remove central banker's discretion over portfolio choice.

Min: if you tax consumption of apples, people consume fewer apples than where marginal benefit = marginal cost. If you tax holding currency, people hold less currency than where marginal benefit = marginal cost. Milton Friedman The Optimum Quantity of Money.

louis: "Moreover, absorbing the idiosyncratic risk of any one market will distort price and so likely distort market micro outcomes. No good."

Suppose a commercial bank, a pension plan, or an individual, holds a risky asset. Does that "distort" price?

Hmm. An asset that is risky for an apple producer to hold may be safe for a banana producer to hold (e.g. an insurance contract against a failure of the banana crop). And that will be different again for what is safe or risky for a money producer to hold.

A safe asset for a money producer to hold would be an asset that gives higher total returns in states of the world where the demand for money falls, and lower total returns in states of the world where the demand for money rises.

Central banks shouldn’t buy bonds or any other private asset because no central bank is entirely independent: that is, all CBs are under political pressure. And politicians always want to paper over cracks rather than tackle fundamental problems: e.g. having the ECB buying Greek bonds at the full nominal value rather than a value that reflects the risk. That kicks the can down the road for a while. Politicians love kicking cans.

During the crisis, were private banks made to pay the “penalty rates” advocated by Walter Bagehot, or was the collateral they offered all “first class” as advocated by Bagehot?

Nick Rowe: "if you tax consumption of apples, people consume fewer apples than where marginal benefit = marginal cost. If you tax holding currency, people hold less currency than where marginal benefit = marginal cost. Milton Friedman The Optimum Quantity of Money."

A: Sounds like Silvio Gesell. ;)

B: Sounds like a fallacy of composition. :(

C: "Tax" is being used with different meanings in the two statements. It is metaphorical in the second statement, right?

If a govt gave money to people because they have a higher propensity to spend it, thereby supporting basic demand, there are complaints, even when it is well understood that it would be good for the economy to get basic demand supported.

In comparison, if a thing called a central bank, creates money and gives it to someone with a higher propensity to spend while entering on its books some financial position, we get a long thread to discuss what it means.

In times when an economy would be advantaged by increased basic demand coming from those with a higher propensity to spend the dollars they have, then we could have do both doing it (the govt adds to its spending and this thing called the central bank gives money to people who will spend it, with bookkeeping to match).

If this central bank is buying safe assets, giving the person safe currency, it is giving the money to a class of people who have a lower propensity to spend; it seems to me, it is people who are already safe and are just trading for safety sake. This makes me think of the upper deck of the Titanic, lot of chairs to shift around for that class of people.

In contrast, look at how the US FRB still purchases MBS because it wants more money into this sector, and in so doing, it is underwriting classes of people who need lower mortgage rates in order to 'buy' a home.

If the point is that you want to do something to support aggregate demand or sector-specific demand, then central banks should give money to those who fit the definition. Oh, and by the way, this 'bank' then records something on the other side of the books where the position is held until term while facing the need to write off some positions, to net to something. The point is about supporting aggregate demand - is that not the point? Isn't the point that you need to recognize that shifting the deck chair on the upper deck isn't much (there are other meanings to this too).

The point isn't "industrial policy" or "bookkeeping profits" for govt institutions - is it?

What is the point: aggregate demand, profit seeking portfolio maximization, industrial policy, something else?

Central banks should avoid doing things that subsidize wasteful, irresponsible or dishonest private sector investment schemes. If the central bank provides an offloading market for the assets generated by such schemes, then they are in the business of incentivizing wasteful, irresponsible or dishonest private investment.

In accordance with the first aim, central banks should make diligent efforts to accurately assess the real quality and expected returns from the assets they buy, and to pay prices for those assets that reflect that real quality. If they believe that the current market price for the asset overstates its real quality, they shouldn't buy it at that market price.

Volatility in the income generated by the central bank's asset portfolio means volatility in the disbursements central banks make to their operating governments, which could in some cases cause political difficulties and occasions for partisan wrangling and foolishness that central banks should (and do) generally try to avoid.

Buying Greek bonds is not in itself a bailout of Greece, no matter how risky they are. If however, Greece were to default on the bonds, then I believe the loss would be born collectively by the EZ governments that are the recipients of the ECB's profits. But that depends on what special programs are or are not in place. For example, I believe the current program sends all profits on Greek bonds directly back to Greece, rather than putting them in the pool of profits that are distributed back to the individual national banks (and governments). That makes Greek debt purchased by the ECB interest free.

Dan: "In accordance with the first aim, central banks should make diligent efforts to accurately assess the real quality and expected returns from the assets they buy, and to pay prices for those assets that reflect that real quality."

That's the fundamentalist (e.g. Graham and Dodds) theory of value investing. You look for stocks (and bonds) that you think are undervalued by the market, and buy them. But if you think that other value investors are already doing just that, and you aren't smart enough to beat the market, you could do just as well by throwing darts to pick stocks, or just buying the index. My old post on EMH

Nick, my point isn't about the central bank looking for value. It is only that the central bank should do everything it can to avoid subsidizing market stupidity.

A central bank should old as much in illiquid risky assets as results from buying as much in illiquid and risky assets as necessary to hit its NGDP target. Ceteris paribus the more fiscal policy is focused on "austerity" (not investing in activities with positive NPV at long term bow wowing cost) the more illiquid risky private sector assets will be required.

JKH, let's go over who gets the fed's (central bank) profits.

Currency holders get a physical asset with 0% interest.

Central bank reserve (demand deposits of the fed) holders get whatever interest rate the fed chooses.

The preferred stock holders get up to 6%.

The federal gov't gets the rest. Is that right?

Why shouldn't police departments lend to small businesses? Why shouldn't the post office rent out real estate?

Well, institutions have charters which are areas of responsibility and expertise, as well legal boundaries on behavior.

So we can ask why the central bank would be good at deciding whether to give me a personal loan or not, or whether to invest in my exciting new blogging venture. It doesn't seem to me that the central bank has a lot of experience to make this type of decision. If the central bank gave me a personal loan, then other people would also want a personal loan, and firms would want a personal loan as well. Would the central bank give *everyone* a loan -- demand determined monetary policy? -- or would it ration who got loans until it's loan target was met -- those well connected to the central bank get loans. Legislatures like the ability to distribute pork and do not want to share it with other, non-elected, institutions.

Also, the decision to invest in risky assets means that someone has to be prepared to bear the loss. As the Treasury receives all central bank interest income, a series of bad investment decisions means that the income is zero or negative, so it makes sense that Treasury would need to give permission to the central bank to do this, which would require regulation of the central bank just as private banks are regulated, so ultimately the big risk is loss of central bank independence as soon as the central bank starts conducting fiscal policy.

If the central bank purchases only government guaranteed debt, then it is not exposing the Treasury or the taxpayers to any additional risk, and so the central bank can act independently.

I think Nick is asking the question because of limits to monetary policy at he zero bound. Under current restrictions, the central bank cannot "print money" in unlimited amounts, since it must use that money to buy government guaranteed debt, which is finite, and the decision of the government to guarantee a certain type of debt is a fiscal decision, not a monetary decision. That finite amount is a limit on monetary policy. Nick does not believe that this limit needs to exist, whereas I do. E.g. fiscal policy is limited by the total amount of debt that the private sector is willing to hold whereas monetary policy is limited by the total amount of debt the government is willing to guarantee. To blur this distinction is to eliminate central bank independence, or to combine the central bank with the Treasury + Legislature into a single government branch headed by a bunch of crazy econ geeks, whereas us non-econ geeks prefer separation of powers and don't want our central bank guaranteeing private sector debt, even if this means we need to acknowledge limits on monetary policy.

Nick Rowe: "If you tax holding currency, people hold less currency than where marginal benefit = marginal cost. Milton Friedman The Optimum Quantity of Money."

I checked out what Friedman said. (At least what others say he said. I didn't dig up the book at the library.) IIUC, his argument is that people (irrationally) do not hold enough currency. (The old guys who buried their money in a sock in the back yard were on to something. Hoocuddanoed? ;)) Therefore the thing to do is to reduce the amount of currency in the economy until people are holding just the right amount. (That is always possible, even if it means doing without currency entirely.) Somehow that is deemed to be welfare maximizing. Right! ;)

It's also interesting to note that there are example of government institutions buying pools of private sector assets -- sovereign wealth funds come to mind, but also various state run pension funds. These types of funds need to navigate some thorny issues:

* variability in returns
* political pressures to use the ownership shares to put pressure on/reform corporate management (which is what shareholders are *supposed to do*)
* political/social concerns (e.g. do you divest from Isreal, from CO2 intensive industries)
* Economic concerns -- In bad times, Chinese State Owned Enterprises are encouraged to invest more, for example.
* Opportunities for abuse/control -- If a particular firm is viewed as a bad actor, do you refuse to roll over its bonds? Do you use your ownership state to kick out the current board of directors?

Ownership is control, so when the government owns a large chunk of the "private" sector, it exerts a large amount of control over the private sector. In a private ownership economy, we circumscribe that control through a political process. It is never just an economic or 'monetary' process.

I don't think this is a technical glitch, but it shows the limitations of thinking about the economy as a homogenous system of equations divided by a single "p", which is determined by a single "M" controlled by the government. When the "M" is controlled by both the government and the private sector, you cannot have control over all of M unless you interfere with the private sector in some pretty fundamental ways. Those actions, under our political arrangements, fall under the rubric of "fiscal policy" and "business regulation", and we have checks and balances in place that, if the central bank were to follow, it would stop being a purely economic agent, and would become a political agent, albeit an unelected one.

Not done yet!

We can look at the ECB, which is in a weird of position of not having a European Treasury to back it, and therefore (IMO) not really being a proper central bank in the first place. Because it has no European Treasury, it's not in a position to purchase bonds guaranteed by the European Treasury, and so it *must* purchase *only* risky debt. It scours the European union and directly purchases the debts of private banks, private firms, and other european governments (their debt being risky because they don't have their own currency).

Now we know that the ECB does purchase Greek/Irish debt, and of course we also know that it would never abuse this position to demand political concessions from the borrowers, right? It would never interfere with the internal politics of these nations.

Because creditors never do that.

So what could go wrong if central banks started purchasing large swathes of private sector assets? It's just a monetary operation, right?

Getting back to the realm of real central banks (e.g. ones with a Treasury backing them) like in Canada, the UK, or U.S., there is a very clear distinction between monetary and fiscal areas of responsibility, at least as codified by law, and just looking at the politicized bully that is the ECB versus the mostly technocratic behavior of real central banks is a clear reason why we should have the former institutional arrangements rather than the latter.

Min: You don't understand it correctly.

Start here (second diagram): http://en.wikipedia.org/wiki/Deadweight_loss

rjs: define a "standard bundle" of Eurozone government debt, where bonds are weighted by (say) the populations. The ECB decides how many bundles to buy, at auction. I think that would be better than what they have now.

Gold is a very risky asset, but central banks used to hold lots of gold, and some still do. Same for foreign exchange reserves. And long term government bonds are risky assets too, but all(?) central banks hold those. Is it really *risk* you are worried about? And how do you define "risk" (risk relative to *what*, because unindexed government bonds are risky, in real terms).

"I think Nick is asking the question because of limits to monetary policy at he zero bound."

You think wrong. And it doesn't matter anyway. Stop looking for conspiratorial motives like some crazed lefty. It pisses me off. It's the job of interlektuals like me to ask the dumb questions that don't get asked. Because we can learn from asking them. Greek bonds were what originally got me thinking, but that doesn't mean I'm really a closet Greek seeking a bailout. Plus my old post http://worthwhile.typepad.com/worthwhile_canadian_initi/2008/12/central-banks-should-bet-on-recovery-literally.html

What we should be looking for is an asset that is "risky" in exactly the right way, so that central banks are forced to print more money to cover their losses exactly when we want them to print more money. Like an asset that falls in value when the price level falls and unemployment rises, and rises in value when the price level rises and unemployment falls. Or an asset whose returns are negatively indexed to NGDP. Government bonds have exactly the wrong sort of risk.

Nick: "Suppose a commercial bank, a pension plan, or an individual, holds a risky asset. Does that "distort" price?"

All the above hold risky assets, because someone must hold each asset. We assume that the commercial banks, pension plans, and individuals acting in their own interest get the relative valuations of these assets right, sending signals to the market. If they are willing to hold more risk in general or revise their views of specific countries or industries, the change in the price of risk will have an effect on the amount of risk supplied.

Maybe I was wrong and we need a better model. If individuals' risk-aversion leads to total investment spending being more conservative than is optimal, then the ideal risk premium is zero, and the CB should do what it can to structurally reduce the risk premium to get it closer to zero.
Maybe the central bank should be in the business of writing options on major market indicies.

louis: see my next post. It should be writing options on deviations of inflation/NGDP from target.

Nick Rowe: "Min: You don't understand it correctly.

Start here (second diagram): http://en.wikipedia.org/wiki/Deadweight_loss"

Thanks. :) I was acquainted with the concept of deadweight loss, but that article had some information that was new to me.

However, deadweight loss is not my concern. I do not have time in the near future to go to the library and find Friedman's book. So I cannot address his argument directly. Maybe I am getting the wrong idea from other sources.

My first concern is the use of the metaphor of a tax, which may or may not be applicable. My second is the impression that important variables are being left out or swept under the rug. (To use a metaphor, myself. ;)) It is as though the optimal number of cookies was being calculated based upon the total capacity of children's stomachs. Other things are important besides the children's desires for cookies.

Oh, yes. My third concern has to do with evidence. Where is the evidence that people are holding less currency than they should? Are they attempting to hold as much as they can by burying it in the back yard, for example?

Min: take that second diagram in Wiki, put M/P (real stock of currency) on the horizontal axis, and i (the nominal interest rate) on the vertical axis, and draw the Marginal Cost curve horizontal at roughly zero (ignoring the central bank's printing and admin costs). Empirically the money demand curve does slope down (we hold more money when the opportunity cost of holding it falls), so if the monopoly central bank sets i above marginal cost, we get a deadweight cost triangle.

@ Nick Rowe

Thanks. You have adapted Friedman's argument to modern practice, I guess. But my second concern, that of leaving out important variables, remains. Also:

Nick Rowe: "If you tax holding currency, people hold less currency than where marginal benefit = marginal cost."

By "tax holding currency" you mean have inflation, right? Empirically, during periods of inflation, people tend not to hold on to money. I suppose that you identify when marginal benefit = marginal cost, people are indifferent to holding currency. But then they try to get rid of it when they hold **more** currency than the Goldilocks amount, not when they hold less.

As for Gesell, his money did have an explicit tax on holding currency. But that was considered a welfare enhancement, not a welfare cost. We cannot simply say that the optimal amount of money in an economy is that amount such that people, on average, are indifferent to holding it or not.


"Why shouldn't central banks buy risky (and illiquid) assets?"

When a central bank actively buys and sells assets, do they become less risky and more liquid - I think so. Can a central bank eliminate the risk associated with any asset - I think to a point it can. Obviously people and may goods have fixed lifetimes. A steel hammer left to rust will turn to dust and scatter in the wind.

And so maybe the question should be - should any asset bear risk?

I know Nick thinks any seignorage will EVENTUALLY be spent by government, even though, at its most basic, it's money that will NOT be spent by government. Buy peoples tax bills (or a portion of). payroll taxes? How do you suck money out? If the default rate is 100% (for illustrative purposes) that wouldn't be hard at all. Plus, it avoids the fact that seignorage is impossible at the ZLB.

oops. wrong blog post. meant for what assets risky assets cb should own post


I'm confused by the harshness of your response here. I point out that you are hiding fiscal policy in monetary policy and you respond with some ad hominems followed by "...so that central banks are forced to print more money to cover their losses". What I'm saying is pretty orthodox and, (to me), should be clear to most readers. Perhaps I'm not clear enough.

Let's try again.

Your proposal is (almost) isomorphic to the following:

Suppose you want the CB to purchase 1/3 of all private sector assets. To do this legally, you can:

1. Have the Treasury make a lump sum payment of 1/3 of the value of the asset to the owners.
2. The payment is financed by selling government debt.
3. The CB buys the government debt with QE.
4. The Treasury institutes a capital income tax equal to 1/3 of all capital income.
5. Proceeds of the tax are used to redeem the government bonds held by the CB.

This is *almost* what you want, except that

- Because we tax the capital income rather than buying the asset outright, we avoid the ethical problems of bureaucrats sitting on executive boards -- e.g. we limit govt. control of the private sector to receiving the dividend/coupon, and not running the company. This is a good thing.

- It is legal.

Note that here the fiscal and monetary components are clearly defined. It's not something that the CB can do on its own, and for good reason.

This should be enough to convince the casual reader that your proposal has embedded in it a fiscal component -- something I see repeatedly here. Don't get angry. That does not make me a deranged lefty or whatever.

Now we can have a more fruitful discussion of whether this would work to stimulate the economy. I'd be in favor of it if it had redistribution as well -- e.g. don't give the lump sum payment to holders of the asset, but distribute it progressively.

I really like the capital income tax, and have advocated for it before. If you believe that the market clearing interest rate is negative, then this can be raised to zero by a high enough capital income tax. I don't understand why there are repeated calls for lower real wages if we really believe there are excess savings demands.

Oh, and if you only want the CB to buy asset class X, then make the lump sum payments to holders of that asset class and only tax the capital income of that class. This generalizes to whatever you want to do. It's still a combination of fiscal and monetary policy, though.

rjs: you accuse me of semantic legerdemain by "hiding fiscal in monetary policy", and miss the point of this post, and wonder why I get pissed?

Not every question is: "should we use monetary or fiscal policy to stimulate the economy?" And this one isn't. This is not a post about monetarists vs fiscalists. And I don't give a damn about whether you call it "fiscal policy". That is a mere semantic argument.

We have these things called "central banks" that, in most countries, are owned by governments but are to a greater or lesser extent treated as independent of the rest of government. They have their own set of books, with their own assets and liabilities.

So we might want to ask some basic questions like:
Why do central banks need assets?
If they do need assets, what sort of assets should those be? Should they only include safe and liquid assets? Why?

And "they must only hold government bonds, because if they held anything else they would be doing fiscal policy not monetary policy! And that would be against the law!" is not a good answer to those questions. It's not an answer at all. It's pure semantics. And then you accuse *me* of playing semantics by "hiding fiscal in monetary policy".

"Because it would be risky if they held anything other than government bonds!" is a better answer.

Yes, if we consolidate the central bank's asset sheet with the rest of the government, we cannot see the difference between the central bank holding risky assets and the rest of the government holding risky assets. But for some reason central banks have their own set of books, and are run at least semi-independently of the rest of the government. If you think that independence is a pure fiction, then it doesn't matter at all whether it is central banks or the rest of government that holds the risky assets. So we still don't have an answer to my question.

Plus, there are some central banks, like the ECB, that stand alone, and don't really have a federal government.

"Plus, there are some central banks, like the ECB, that stand alone, and don't really have a federal government."

Yup, this is the point. The ECB is not a real central bank because it cannot buy risk-free assets. There are no risk-free assets denominated in Euros.

So the ECB conducts fiscal policy (because it can't avoid it), and as a result it's telling Greek Women that they need to retire later, it's telling Greek Unions that they need to scale back their wages, it's telling the Irish government what kinds of surpluses it needs to run. You know, that kinda sounds like fiscal policy, doesn't it?

Now these are things that creditors do all the time. These are also things that *Governments* do all the time. But governments have to conduct transfers, and -- we hope -- they exert their authority with some form of democratic accountability.

Now you -- not me -- want this new fangled thing called 'Central Bank Independence' -- in which a bunch of unelected economists get to dictate things to the rest of the economy.

Fine, as long as the legislature tightly controls what they are and are not allowed to do.

They are not allowed to conduct transfers, which means that they can't do fiscal policy. This is a good thing, because they can't set the terms of the transfer, either. Because they can't do fiscal policy, the electorate is OK with them being unaccountable and unelected.

As soon as they start doing fiscal policy, as soon as they become a quasi fiscal-monetary agent, then they either lose their independence to elected bodies, or they (God help us) become like the ECB.

You mentioned forex pegs and buying gold (really a forex peg when CBs invested in gold). Wanna know why none of those CBs were independent? It's because they bought those risky assets. You can't do all three: maintain a peg, have an independent monetary policy, and have an independent fiscal policy -- for purposes of pure logic/accounting.

For purposes of living in a democratic society, you cannot have unelected central banks conducting fiscal policy (which is why the ECB is so hated and europe is so messed up). That's why the CB can't buy risky assets.


"For purposes of living in a democratic society, you cannot have unelected central banks conducting fiscal policy (which is why the ECB is so hated and europe is so messed up). That's why the CB can't buy risky assets."

But you can have a fiscal authority sell risk assets while maintaining democratic accountability. In fact, that could cure a lot of what is wrong in Europe.

Sure, if the EMU wants to create a single government or otherwise unify, then should do that with a european treasury, with a euro-income tax and a european ECB. Then all "bailout" functionality shifts out of the ECB and into the Treasury side of things, and the job of the ECB is to set euro-wide interest rates, rather than determine how much Greek janitors should get paid.

I blame the current faults of the EMU squarely at the feet of the monetarists who really thought all you needed for cyclical demand management was an interest rate policy as set by the central bank. So they created a monetary union with a central bank and no treasury. But of course fiscal policy is always central in a crisis and so fiscal policy sneaks back into the only institution available, the central bank, creating a type of franken-bank monster. The whole notion of the ECB demanding concessions about how much Greek janitors get paid before it stops providing liquidity to Greek banks is so offensive you want to kill the monster but its the only institution there to conduct the fiscal transfers necessary. If this is an example of anything, it's a cautionary tale about what happens when you don't plan for fiscal policy in times of crisis.


No. Creating a single government to unify the members of the EMU would create more problems than it solves - different languages, social norms, education systems, legal systems would need to be melded together. Each member of the EMU has a Treasury that can sell risk assets to it's own populace and that really is enough.


If we decide the central bank should invest in risky assets as part of its normal operation, what's the stopping rule?

Say the CB starts buying an index fund in addition to government debt. What stops us from repeating your argument and asking "well why shouldn't the central bank buy individual stocks?". And so on, until we reach the point of the CB making individual home loans.

One stopping rule you've alluded to is the idea that index funds are a sort of "neutral" investment. But there is at least some basic level of economic calculation involved in choosing a fund. It's not a lot on a per-$ basis, but multiply that over the size of a CB balance sheet and you're talking real distortions.


There are collateralization requirements in the US, of course -- interestingly, they claim assets like gold and SDRs as backing physical currency instead of all-Treasury /mortgage bonds. Line 7 http://www.federalreserve.gov/releases/h41/Current/ Might this just possibly suggest that gold and SDRs are senior assets?

But at the end of the day, there is no deliverable for the FRN liabilities, and the Fed has never shrunk its balance sheet by any noticeable level — i.e. engaged in net selling -- so who cares what they have on the balance sheet?

At worst, they could maintain a liquid high-quality trading balance of some 10% of assets. The other 90% of the Fed's balance sheet could be a 1oz gold coin that they value at $3.7 trillion.

Hey, they still value the gold at $42, why not just call it whatever price they like?

I had been under the impression the CB buys risky and suddenly illiquid assets mostly to stop the vicious cycle of mark to market losses, margin calls, forced selling that comes with a panic. Especially when credit default swaps on trading banks threaten contagion. During the '08 panic, dealers could rationalize distresses prices of CDOs and derivatives thereon only by forcing literally impossible correlations into their models. But they could not buy these assets and keep their jobs.
Every party has to de-lever. Every party is marked to market. Except the Fed. So in some ways a shrewd fed can pick off the market, since the regular market bid is zero.
Maybe this is the last resort/ stabilizing function Nick mentioned above.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad