« How to find a topic for an economics research essay | Main | Taxes and the value of paper money »

Comments

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

Is the problem here defining a frame of reference?

In normal times, actual price inflation or broad monetary inflation (let's be clear about what kind of inflation we are speaking of) reduced the value of its liabilities more than it does its assets.

But a default by the government reduced the central bank's assets most of all while leaving its liabilities alone.

Now that I have written that, that difference seems to be the key to what will happen and what the central bank will do about it. But this really is hazy, crazy stuff. Why can't we get back to a normal world?

As a practical matter it's not likely to impact policy in the US. I don't know about other countries. The irony is that much of the damage done to balance sheets was caused by the tight money policies in the US, eurozone and Japan, which reduced nominal GDP and asset values. These policies also reduced real interest rates, so central banks earn less income on their assets.

Money market mutual funds don’t invest in the liabilities of their owners, or certainly not exclusively.

When the central bank invests exclusively in treasury liabilities, it just becomes a conduit for deficit financing. The interest on its assets is economically irrelevant. So is its capital position.

The only material question is the cost of its own liabilities as part of the mix of the effective external cost of deficit financing.

The worst case for its stated profits and capital is that where revenue from its assets is zero, such as might be the case at hyper-zero bound. Then the central bank can only run negative profit, at least due to operating costs. And it can only run down its capital position to zero and negative.

But the only relevant factor is the profit or loss. A loss simply adds to effective external deficit costs. The capital position doesn’t matter from an economic perspective, because in the same hyper-zero bound case the government can just swap a capital injection for a zero interest government liability. And this entry per se has no effect on the deficit. The capital transaction is a flow of funds transaction, not an income or deficit affecting transaction.

The only economic issue is the contribution of the central bank (as a funding vehicle) to the cost of the deficit.

And the economic value of monetary liabilities is their book value, because the central bank is in a unique position to sustain any positive or negative capital position the government desires from a bookkeeping perspective. The capital position is economically irrelevant, so there’s no need for the value of monetary liabilities to be affected at all.

Are we talking about central banks in the abstract or some central bank in particular? The new money the CB creates to buy bonds just goes into reserves during a liquidity trap anyway. Many CBs have the power to simply raise reserve requirements in order to prevent this money from being released; they don't need to buy the money.

Anyway, in the case of the US, we are talking about default as in late payment, not default as in complete repudiation. Once the economy recovers and the CB wants to remove base money, the bonds will no longer be in default and will recover their value. So the CB will be theoretically insolvent when it wants to create money but this insolvency can never be realized. How then can the prediction of your theory be observed?

Determinant: "In normal times, actual price inflation or broad monetary inflation (let's be clear about what kind of inflation we are speaking of) reduced the value of its liabilities more than it does its assets."

Assume the Bank of Canada's assets are 100% Canadian dollar bonds. I would say that an increase of the *price level* would have the same effect on the real value of its assets and liabilities. But an increase in the *expected inflation rate* would reduce the real value of its assets more than the real value of its liabilities, if some of those bonds are longish bonds, because long bond prices fall when an increase in expected inflation increases nominal interest rates.

"But this really is hazy, crazy stuff. Why can't we get back to a normal world?"

I'm feeling that way too. The world is not supposed to be posing exciting questions for macroeconomists to try to answer.

Scott: "The irony is that much of the damage done to balance sheets was caused by the tight money policies in the US, eurozone and Japan, which reduced nominal GDP and asset values."

Aha yes! That's a big difference between micro MMMFs and macro central banks I hadn't mentioned!

"These policies also reduced real interest rates, so central banks earn less income on their assets."

I think you meant *nominal* interest rates? (Assuming cash pays zero nominal?) Or do you count the negative real interest on currency as income, plus the positive real interest on assets? Which is a different but a consistent way of thinking about it.

JKH: "Money market mutual funds don’t invest in the liabilities of their owners, or certainly not exclusively."

Lovely!

"When the central bank invests exclusively in treasury liabilities, it just becomes a conduit for deficit financing. The interest on its assets is economically irrelevant. So is its capital position."

If a central bank is just a branch of the Ministry of Finance, and has no independent decision-making authority, I think that's the correct way of looking at it. There is really no point in not simply consolidating the Bank's balance sheet into the government's overall balance sheet. Simply cancel the government bonds held by the central bank, and it all washes out. In that case, it's hard to see how a government default could fail to affect inflation, and expected inflation. Inflating the currency is just one of the ways in which the government can lower the value of its liabilities when it is insolvent at the existing price level.

But when the central bank has some degree of independence, it's different. We want to know if the central bank has the ability to do what it wants to do. And that *may*, or may not, depend on the value of its assets.

Phil: "The new money the CB creates to buy bonds just goes into reserves during a liquidity trap anyway."

If those reserves pay 0% interest, they are the same as currency from the POV of the central bank's seigniorage profits.

"Many CBs have the power to simply raise reserve requirements in order to prevent this money from being released; they don't need to buy the money."

Damn! I forgot about that.

If "default" simply means "late payment" (but of the same PV, as in a compulsory rollover), I don't think it would matter much for central banks. It might matter for markets that need liquidity? And that might indirectly affect central banks? Dunno.

Ok Nick, the question seems to be what happens to the Central Bank's equity position and how does the CB respond to changes in it. For clarity, Equity = Assets - Liabilities. If a shock decreases assets more than liabilities, the CB's equity falls. If a shock decreases liabilities more than assets, equity increases.

What are the different policy responses to falling CB equity vs. rising CB equity? If we're having a balance sheet problem that seems to the best way to clarify the problem since equity is summary of the balance sheet position.

I don't think central bank independence is the issue either. The capital position is economically irrelevant, whether or not that happens. Either way, if a treasury default results in a central bank capital shortfall, that is easily repaired by internal bookkeeping as described. There may be an operational question as to who's calling the shots on monetary policy. But the consequent cost of central bank liabilities and its contribution to the cost of financing the deficit remains the relevant economic issue. And the value of monetary liabilities shouldn't be affected, given the economic irrelevance of the capital position - whether or not the central bank has lost its independence.

Nick:

That effect that you called a kind of multiplier effect (I call it 'inflationary feedback') is on page 8 of my paper "There's No Such Thing as Fiat Money". Here's the relevant passage:

"The 100 ounces of silver deposited with the banker can be called physical backing for the dollars, while the farmer’s IOU, being denominated in dollars, can be called financial backing. Define E as the exchange value of the dollar (oz./$). Since assets (100 oz. plus the $200 IOU) must equal liabilities (300 paper dollars), it must be true that 100+200E=300E, or E=1 oz./$. If the bank did not have adequate backing for its dollars, inflation would result. For example, if the bank were robbed of 20 oz. of its silver, then the above equation becomes 80+200E=300E, or E=0.8 oz./$. If the bank had backed its dollars with 300 ounces of silver and held no IOUs, then the loss of 20 ounces would yield the equation: 280=300E, or E=0.933 oz./$. Note that a higher level of physical backing (as opposed to financial backing) makes the value of the dollar less sensitive to changes in the bank’s assets."

Mike: post updated. Thanks.

Hi, Nick. So many interesting posts. You are on a roll!

"But I think it quite possible that increased risk of sovereign default may reduce the expected future real value of the central bank's liabilities. It increases expected inflation, in other words."

If I understand your argument, it would increase expected inflation in a "normal economy", not in an economy that is in a liquidity trap. As you say:

"A large enough loss in the value of a central bank's assets will make it impossible for the central bank to redeem enough money to keep inflation on target when the demand for its money falls."

If the demand for money remains high (desired savings exceed desired investment), there is no need for the central bank to redeem money in the first place.

So your post largely applies to a normal economy that is not in a liquidity trap. Is this correct?

The question in the first paragraph of Nick’s article was what happens to the value of a currency when the value of govt securities drops.

Let’s assume a complete default, i.e. govt says it won’t reimburse bond holders on maturity of the bonds. This has little bearing on the value of the currency, and for two reasons. Currencies have value, first because people and businesses far prefer a currency, any old currency, rather than have to barter. Second, the reason the state’s currency is normally the dominant currency derives from the state’s power to tax.

Neither of these factors are affected or diminished by a govt default. So no effect there.

In a closed economy there’d be a reduced trust in government as a whole (i.e. including the central bank) and thus a flight from storing wealth in the form of money, i.e. a flight to gold, property, etc. That would be stimulatory / inflationary. On the other hand bond holders would experience a loss of net wealth, i.e. their net financial assets would fall below their desired level. To that extent, they’d save more, which would have a deflationary effect. But they are hardly likely to want to store wealth in the form of the relevant government's currency or bonds. So the former effect would predominate.

As to an open economy, there’d be a reduced trust in the relevant government (e.g. China’s trust in the US govt has declined recently – can’t blame them in view of the childish pranks on Capitol Hill). That would cause a flight from the relevant currency by international operators, which would reduce the purchasing power of the currency both internationally and domestically, particularly the former.

Kien: "So your post largely applies to a normal economy that is not in a liquidity trap. Is this correct?"

For a central bank that is truly independent, I think it is more the other way round. In normal times a central bank has lots of monopoly profits, so if it makes losses on its balance sheet it can rebuild its assets quickly. Plus, it doesn't need much assets anyway, because it is unlikely to want to reduce the money supply much in the near future. But in a liquidity trap it is earning no monopoly profits, and is likely to want to reduce the money supply by a large amount in the near future when the economy escapes the liquidity trap.

Ralph: "Currencies have value, first because people and businesses far prefer a currency, any old currency, rather than have to barter. Second, the reason the state’s currency is normally the dominant currency derives from the state’s power to tax."

Your first reason is the standard quantity-theoretic reason why currency has value. Your second reason is more chartalist/MMT, but can nevertheless be understood, in part, within the quantity-theoretic tradition. Like this:

Yes, there is a demand to use, and therefore hold (for temporary periods) money as a medium of exchange. And that's why even intrinsically worthless paper money can have value. But that demand curve slopes down. The willingness to hold it (the amount people are willing to hold on average -- how quickly they will want to get rid of it) depends on the cost of holding it. And that depends on how quickly it is depreciating (inflation). And that depends in part on how quickly the stock of money is growing over time. And that *may* depend in part on taxes that withdraw money from circulation.

The value of money depends on the current stock of money and the demand for that stock of money. The demand for that stock of money depends on the rate of inflation, which depends on how quickly the stock of money is growing over time, which may depend on how high or low taxes are.

For example, if a government were printing and spending money, but had no taxes, the stock of money would be growing very quickly over time, so inflation would be high, so the demand for money would be low so the value of money would be low so the price level would be high.

So, two otherwise identical economies, with exactly the same stock of money today, could have very different price levels today, if one has low taxes and a quickly-growing stock of money and the other had high taxes and a slowly growing (or even falling) stock of money.

But, if the government can also issue bonds, that people are willing to hold because they pay interest and not because they are useful for shopping, then the simple link between taxes, spending, and the growth of the stock of money gets broken.

The above is how a quantity-theorist would incorporate the Chartalist/MMT/Fiscal Theory of the Price Level view -- that the value of money depends on taxes -- into the quantity theory. And there is nothing new in the above, by the way. What I wonder about is whether MMTers understand that monetarists incorporate taxes into the analysis of the value of money that way?

Monetarism is money stock based – it considers taxes marginally to the degree that taxes affect the stock of money - very marginally. Money is the swing factor for aggregate demand.

MMT is income flow based – it incorporates taxes centrally to the degree that taxes affect income - very centrally. Income and satiation of saving desire are the swing factors for aggregate demand. Money stock as conventionally defined is ignored. Money stock to be useful must be defined transformatively according to MMT (e.g. "net financial assets").

Taxes are obviously contractionary for both.

Monetarism treats government bonds as contractionary – because they reduce the money stock.

But MMT treats government bonds as expansionary – because they increase non government income compared to money, and are as liquid as money via collateralization (so says MMT).

I don't know if this is relevant information, but the Feds paid the government a record $81 billion in profits in 2010. So as a monopoly MMMF, 2010 was a very good year.

Lets just ask what happened with Argentina. Default and then inflation.

"I think it quite possible that increased risk of sovereign default may reduce the expected future real value of the central bank's liabilities."

This was the source of inflationary expectations in high-inflation Latin American countries. When the C.B.'s assets are in doubt, inflation rises in the presence of falling growth and a large output gap. From an NGDP targeting perspective, this would be a difficult problem.

Are there two types of inflation? One occurs when the economy "overheats"; the other occurs when markets doubt the value of Central Bank assets. I see many economists arguing that we cannot have high inflation as long as the "overheating" condition is not met. Further, the nature of "doubting C.B. assets" is that it occurs in non-linear fashion in response to triggers. That is, markets can be relatively complacent until some trigger event, at which time they panic. Economists that point to stable sovereign debt yields as evidence of future stability often get very surprised by the speed at which those same yields spike.

JKH: I'm writing another post on taxes and the value of money, that lays out the basic quantity-theoretic position. Sort of an extended reply to Ralph. Your comment reminded me to add something to that post. Thanks. We can continue that discussion there.

beezer. That is very definitely relevant information. Tends to work against my theory. But maybe this was a one-off gain to the Fed from buying up assets on the cheap during the financial crisis?? And since it has already handed those profits over to Treasury, they don't help its capital position any more?

JimP: good example. I see the Euro as being at risk of doing something very like Argentina. But bigger, and more complicated because there are multiple potential Argentinas.

David. Good comment. I normally only think there is one type of inflation. There's an excess supply of money and hence an excess demand for goods. And these are just two sides of the same coin. But maybe in the cases you are talking about the transmission mechanism worked primarily via the exchange rate, so the excess demand for goods was in the form of an increase in import costs rather than abnormally high domestic demand. In general equilibrium, we can't really distinguish between demand pull and cost push. An excess supply of money causes both input and output prices and prices of foreign exchange to rise. Which one comes first temporally depends on adjustment speeds. Exchange rates usually move quickly.

And in Argentina the default investment hedge was stocks - or real estate. Anything except the currency. And the dollar is not doing all that well now either.

Nick,

I agree, except that I think foreign currencies are just one type of hedge that actors use when they come to doubt C.B. assets. Forward buying of commodities, selling the long end of the yield curve, and forward purchasing of inventories are all just as valid as inflation hedges. It is important to recognize that a currency devaluation was a symptom and not a cause of inflationary expectations in Latin America. The cause is actors' desire to hedge against future inflation, which causes an acceleration in current inflation, which causes future inflation expectations to rise further.

Why is this important? Many economists believe that a large dollar devaluation would not cause high inflation in the U.S. given the small size of the external sector. Thus, "Latin America" type inflation is supposedly improbable here. This misses the point that hedging, not devaluation, is the key dynamic.

JimP, David, and me, are all on the same page.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad