Good and bad currency wars

Canadian Finance Minister Jim Flaherty is concerned that Switzerland's actions to depreciate the Swiss Franc against the Euro could lead to a currency war. But currency wars don't have to be bad; they could be good. It depends on how they are fought. If fought the right way, an invisible hand may lead to a good outcome that was unintended by any of those fighting the war.

Just to illustrate what I mean, let's first start with a simple example of the Gold Standard. This would help us understand currency wars of the 1930's.

Under the Gold Standard, every country pegs the value of its currency against gold. And suppose we start out in equilibrium, where every country believes it has sufficient gold reserves to enable it to cover foreseeable circumstances where it might need to sell gold in exchange for its currency at that price of gold.

Then something bad happens that causes every country to want to double its gold reserves. What happens next?

1. There might be a bad currency war. Each country raises interest rates and cuts its spending to increase capital account inflows and net exports to attract gold from abroad. In terms of gold, it's a zero sum game. One country's gain of gold is another country's loss of gold. From a wider context, it's a negative sum game, because the world will suffer recession and deflation.

2. Or, there might be a good currency war. Each country raises the price of gold to devalue its currency to increase net exports to attract gold from abroad.

Why is that second war a good currency war? It's still a zero sum game, isn't it? One country's net exports are another country's net imports? One country's gain of gold is another country's loss of gold?

In terms of gold, it truly is a zero sum game. But from a wider context, it's a positive sum game. Step back and think how the problem could be cured, at a global level. If each country wants to hold double the amount of gold reserves, it can't be done. There's only so much gold in the world. You can't double the tonnes of gold in existence. But you can double the value of those tonnes of gold in terms of money. If every country doubled the price of gold (halved the value of its currency against gold), that would get the world back to equilibrium. One tonne of gold reserves would now be able to do the same job that two tonnes of gold reserves would have been able to do at the old exchange rates.

And the good currency war, where each country tries to gain gold reserves by raising its own price of gold (devaluing its own currency), leads us exactly to that global cure. All it needs is for every country to wage war enthusiastically and keep on retaliating until each has its own desired doubled value of gold reserves.

A simple story of the 1930's is that it started out as a bad currency war. Each country tightened its monetary policy, and that lead to the Depression. And then it eventually turned into a good currency war. The Depression slowly ended as countries devalued against gold and thereby loosened their monetary policies.

That was the gold standard. It's different today. The problem isn't an international shortage of gold reserves; its an international shortage of monetary demand for goods. And we can't print gold, but we can print paper money. And each country wants to expand the monetary demand for its own goods, in part by devaluing its exchange rate to increase net exports. And net exports are a zero sum game. But just as under the gold standard example, there are two different ways to fight the war: a bad way and a good way.

1. The bad way to fight a currency war is to devalue your own currency by making other countries' currency more scarce. Your country might gain, since your net exports might increase; but the world as a whole loses monetary demand for goods.

2. The good way to fight a currency war is to devalue your own currency by making your own currency more plentiful. Your country will gain, since your net exports will increase; but the world as a whole gains monetary demand for goods.

If countries fight a currency war by loosening their own monetary policy, rather than by tightening other countries' monetary policy, it doesn't really matter if each individual country sees most of its gains as coming from increased net exports. That just gives them the extra incentive to be doing something they all ought to be doing anyway.

Here are a couple of hypothetical examples, for a world of two identical countries:

1. Country A devalues by buying B's currency in exchange for A's bonds. Country B retaliates by buying A's currency for B's bonds. The net result is exactly as if each country did an open market sale of bonds. There is less currency in circulation and more bonds. That's a bad currency war. It's a tightening of global monetary policy. It's a negative sum game.

2. Country A devalues by buying B' bonds in exchange for A's currency. Country B retaliates by buying A's bonds in exchange for B's currency. The net result is as if each country did an open market purchase of bonds. There is more currency in circulation and less bonds. That's a good currency war. It's a loosening of global monetary policy. It's a positive sum game. It's global quantitive easing.

Those two examples, being hypothetical, are much simpler than real world cases. But the underlying point remains: if countries are devaluing by loosening their own monetary policy, rather than by tightening others' monetary policy, it's a war we want to break out.

(For more on the Swiss devaluation, see Scott Sumner, David Beckworth, Josh Hendrickson, David Glasner, and Marcus Nunes. What the Swiss experience reinforces, as those bloggers note, is that there is a lot more to monetary policy than just setting an interest rate.)

Comments

[rsj]

Nick,

"Country A devalues by buying B' bonds in exchange for A's currency. Country B retaliates by buying A's bonds in exchange for B's currency. The net result is as if each country did an open market purchase of bonds. There is more currency in circulation and less bonds."

In both cases, there will be neither more nor less currency in circulation as a result of the operation.

Currency in circulation is the paper cash and coins withdrawn by the non-financial sector from their deposit accounts, less the currency deposited by this sector into their deposit accounts.

So how does either CB OMOs, or CB forex purchases force households or firms to turn around and withdraw more or less currency from their deposit accounts? It doesn't.

These withdrawals and deposits are the intentional decisions of deposit holders, as they respond to changes in the price level, changes in payments technologies, or other preferences for transacting with cash rather than by writing checks or using electronic payments.

An increase in the availability and convenience of auto-bill pay systems has a far bigger impact on currency in circulation than a CB forex intervention.

The forex intervention may force the financial intermediaries to quote a different exchange rate, and thus affect trade, and therefore prices, and therefore demand for currency, eventually tricking into more currency withdrawals. But the act of purchasing foreign currency does not cause currency in circulation to increase.

But if all CBs do this, so exchange rates are unchanged, then there will be zero impact on currency in circulation, which is economically irrelevant anyway. I fail to see the transmission mechanism if all CBs do this.

And can we please get out of the commodity money model? That was always a fantasy, but now that we are not even nominally on the gold standard, it's a misleading view of monetary operations -- coming hot on the heels of the "we didn't understand money operations well enough to warn against the euro" post.

[Andrew F]

Is the problem here that the US can't buy Chinese bonds? Or does it not matter what assets the US buys?

Can we be reasonably confident about avoiding stagflation once inflation gets rolling and central banks start to tighten once more?

[Nick Rowe]

rsj: "And can we please get out of the commodity money model? That was always a fantasy, but now that we are not even nominally on the gold standard, it's a misleading view of monetary operations..."

Actually, Canada has had a commodity money for the last 20 years. So have most modern central banks. The commodity is no longer gold. It's the CPI basket of goods. And convertibility is not direct, it's indirect. And it's a crawling peg, at 2% per year, not a fixed peg. We call it "inflation targeting". But it's a commodity money.

You really need to stop thinking of monetary policy as setting interest rates. If nothing else, the Swiss devaluation should teach you that. Did the Swiss cut interest rates to loosen monetary policy and devalue the Swiss Franc? Nope. They just announced a target, and Bingo! That mythical magical confidence fairy did the rest. God only knows how. It doesn't make any sense at all, if the only thing central banks can do is set an overnight rate of interest, and everyone knows monetary policy is tapped out when they hit the ZLB. So there is no way the Swiss could have devalued. Unless they used magic.

[rsj]

Nick, whatever model you use -- it has to be consistent with a basic awareness of the institution.

Central banks cannot buy baskets of goods and services.

Government purchase of output is fiscal policy.

CBs *try* to achieve an inflation target by setting borrowing rates and/or purchasing assets, _hoping_ that the change in borrowing rates and asset prices creates the right amount of inflation.

But it need not, and that is certainly not the same as maintaining a peg.

Btw, I'm waiting for the rush of inflation to take hold in switzerland. I'm sure it will happen any day now...

[Nick Rowe]

Andrew: "Is the problem here that the US can't buy Chinese bonds? Or does it not matter what assets the US buys?"

I think the fact that the US can't buy Chinese bonds means it's a rather lopsided war in that case. If China buys US bonds and/or currency in a sterilised forex intervention (so the supply of yuan does not increase), that's a problem.

"Can we be reasonably confident about avoiding stagflation once inflation gets rolling and central banks start to tighten once more?"

That's the big question. If you think that the problem is structural unemployment, not a shortage of aggregate demand, then loosening monetary policy will only cause inflation. My view is that it's a problem of AD. If AD increases *too* quickly, there might be more inflation than reduced unemployment, because it takes time for firms to hire extra workers and ramp up production to meet the increased demand. But that's the least of our worries right now, globally.

[Nick Rowe]

rsj: "Central banks cannot buy baskets of goods and services."

Under the gold standard, they didn't do a lot of buying and selling gold either. It's called "indirect convertibility". For example, the central bank could peg the price of haircuts without ever buying or selling haircuts -- by adjusting the price of gold. Whenever the price of haircuts starts to rise above target, the central bank lowers the price of gold (sells gold). Whenever the price of haircuts starts to fall below target, the central bank raises the price of gold (buys gold). Or it could use silver. Or copper. Or wheat futures. Or the TSX300. Or bonds, or forex.

Stop thinking about what the central bank is buying or selling. Remember that what matters is that it is buying or selling its own *money*.

Stop thinking about what the central bank is buying or selling *today*. Remember what matters is the central bank's *commitment* to its future strategy.

"Btw, I'm waiting for the rush of inflation to take hold in switzerland. I'm sure it will happen any day now..."

It already has happened. Asset prices jumped instantly on the news. It will take longer for sticky goods prices to respond. And they will respond by *not falling*. because if the Swiss hadn't done what they did, they would have faced deflation.

Like the other "quasi-monetarists", after the Swiss experience I have lost all patience with that closed-minded "horizontalist" perspective, whether it come from MMTers or Neo-Wicksellian mainstreamers. People who can only see the individual trees of the mind-numbing institutional and accounting detail of what they falsely think is how "modern" central banks operate. It's total BS. Read those other posts I linked to. And Scott Sumner's latest.

[jesse]

A good and timely post, I was thinking about the parallels between Canada and Switzerland; there are some interesting ones, especially for exporters in the eastern part of the country who are finding it difficult to compete with a high CAD, as well as the threat capital inflows attempting to find a "safe haven" in a country with better debt ratios.

There are countries who run their currencies as fixed pegs or close to, most notably in Asia after economies got creamed by currency flows. Now they are much more wary of floating exchanges for economies with disparate terms of trade.

Canada, though, is in an interesting spot; could it reasonably peg its currency to the USD? I think, actually, there is some ability to do this as long as its inflation vis a vis the US doesn't diverge. On that front I think that's where the peg would never work for Canada unless the US starts having more of a fiscal policy than its current dysfunctional gridlock. Canada seems more mature here and this could lead to higher inflation if the government's job-producing initiatives start working (whatever they are).

I think Nick what you point out about currency wars and monetary policy is that the implications for inflation depend. When we talk about "currency wars" we also need to talk about "trade wars" because capital flows are an integral part of what this is all about.

[Nick Rowe]

jesse: thanks. We *could* peg the Loonie to the USD (we did in the past), but I think it would be a big mistake. Especially at present. We need to keep an independent monetary policy. The Fed can't even get monetary policy right for the US, let alone get it right for Canada. The Bank of Canada hasn't been perfect, but it could have been a lot worse.

[ObsoleteDogma]

Nick,

Very much agreed. Currency wars is just another name for global QE. But what about capital controls with a currency peg, a la China? Their defense of that peg leads to them pursuing contractionary monetary policy to combat the inflation maintaining the peg and sterilizing inflows would produce. The Fed could, of course, effectively make China abandon the peg if it was determined to do so, but barring that, the dimunition of Chinese demand makes this into the bad kind of currency war, no?

http://obsoletedogma.blogspot.com/2011/09/snb-currency-wars-china-peg.html

[jesse]

"I think it would be a big mistake"

The distinction for the CHF compared to Canada's current situation is that a lot of the flows were financial, not "real", and that was affecting the "real" economy. The Swiss decided their non-bank economy was beggar thy neighbour and was in danger of entering a severe recession if nothing was done. Canada is not in that position, yet, but if people looking for "safety" decide the Swiss aren't worth the hassle, they might start turning to Canada in greater numbers. That is a risk the Bank of Canada and the federal government should be planning for, and I'm reading between the lines on the MoF's recent public statements. So it's a great question, in my view, to ask what Canada could do to shoo off a swarm of locusts.

[marcus nunes]

I called it "multiculturalism". There are those like Plosser who only see MP working through interest rates - so that MP is easy and potentially inflationary, and those like Evans who have a "more opnen mind":
http://thefaintofheart.wordpress.com/2011/09/08/%E2%80%9Cmulticulturalism%E2%80%9D/

[Determinant]

Stagflation has nothing to do with the current situation. This is not the 1970's again, this is the 1930's. Nobody cares and nobody should care about stagflation.

[Patrick]

Nick: Do you think the Fed is too wishy-washy in its communications? The SNB comes across as practically daring someone to test them - "Go ahead punk, make my day!". Also, could/should the Fed say something like "We will buy in unlimited quantities until such time as inflation is X%, and continue until NGDP level returns to trend"?

The SNB experience would seem to suggest it would work.

[Nick Rowe]

Obsolete dogma: Good post. And you made almost all my points yesterday! I *think* you are right about China.

jesse: it's different for Canada because: the US is our biggest trading partner; but when there's a crisis, everybody wants US dollars, not Loonies. Canada is not (yet?) seen as a safe haven, unlike Switzerland.

marcus: "Multiculteralism"? Maybe. Being able to look at monetary policy from more than one point of view. Allowing different conceptual schemes. Framing. Duck/rabbit. Wouldn't that be more like "post-modernism"?

[Nick Rowe]

Determinant: I basically agree. But if AD literally jumped 10% overnight, I would be a little worried. Even with loads of idle resources, it takes time to crank up output. Give it a few months.

Patrick: "Nick: Do you think the Fed is too wishy-washy in its communications?"

Understatement of the year. It doesn't even know what it wants to communicate. Yes, yes, and yes!

[jesse]

"Canada is not (yet?) seen as a safe haven, unlike Switzerland."

My point is that the risk of this should be on the radar of the Ministry of Finance and the Bank of Canada, especially now. I'm trying to work through what could be done given Canada's economic situation of being tied to the US economy but also being a petrodollar. The Swiss trade mostly with the Eurozone so from that POV they are similar to Canada in terms of trade partner diversity. Canada of course has a much larger GDP and population but so does Japan.

[Nick Rowe]

By the way, how does the bank of Canada set the overnight rate of interest? Answer: 99% communications channel. And people at the Bank have told me exactly that. It says it wants the overnight rate at x%, and it goes to x%. It doesn't actually really need to *do* anything, 99% of the time. It's that damned post-modern confidence fairy that does the work.

[jesse]

"It's that damned post-modern confidence fairy that does the work."

This is the so-called central bank bazooka. It isn't a confidence fairy threat, it's a confidence fairy promise. :)

Remember in 2008 when LIBOR and the TED spread went off the charts, the central banks stepped in and cleaned up like a boss.

[K]

Nick:" 99% communications channel."

Of course, that's cause they have the OMO sledgehammer to back it up.

[Jim Sentance]

I was more taken today by his musings about lowering tariffs to counter the problem of cross border price differentials. Now there's a start on a tariff war I could appreciate.

[rsj]

Nick,

Your topic for this post was the argument that forex purchases by SNB cause currency in circulation to increase, which drives up the price level.

This is a simple, testable, and wrong theory. Let's wait and see what happens to SNB currency circulation. What do you think will happen?

It will have be the same result as what happened to currency (and deposits) after the U.S. did QE, or after Japan did Q.E.

Namely, no change.

We've seen this pattern before -- you have a pet paleo-monetarist theory that cannot stand up to observational facts, and instead retreat to faith-based arguments -- "I don't know how it works, but it just *does*, because I assume that it does."

By the way, I *also* believe that the SNB intervention was a good move, and may even be inflationary, but that is because of the current account changes. A low forex rates stimulates exports and hinders imports.

But not if all CBs do it.

And the point of your post was there is an effect _even if all CBs do it, whereas I claim that there is no effect if all CBs do it -- assuming each CB also maintains the same nominal interest rate.

Now as all CBs are not going to do it, we expect it to have a positive effect. That is not evidence in favor your model -- evidence in favor or against your model must be a change in the quantity of currency in circulation, which in your own words, is the cause of the change in prices -- Unless you are going to issue some form of retraction, and then propose an alternate mechanism that would cause prices to increase even if the quantity of currency does not.

But after many interventions we know that the CB has extremely limited capacity to cause the number of non-financial sector deposits or currency to increase as a result of quantitative easing -- even in an environment of positive interest rates. And we also know why -- because the financial sector will for the most part undo whatever changes are made by the CB, as the CB is neither the marginal purchaser nor seller of "money" to the non-financial sector. The only thing left are rates -- forex rates, interest rates -- rates, not quantities.

[rsj]

Btw,

If you want to re-write this post, striking the phrase

"The good way to fight a currency war is to devalue your own currency by making your own currency more plentiful." and related phrases, acknowledging that the amount of currency that households withdraw or deposit is controlled by them in toto, and not by the central bank -- then you can propose some *alternate* channel by which, if all central banks "promise" to cut the forex rate of their own currency in half, then this will result in some form of income boom.

The problem with appealing to the confidence fairy in that case is that households know that the CBs cannot deliver on such a promise -- not if each CB is equally determined.

As an aside, arguing that "CB communications strategies" is what is needed to address mass unemployment has to be the logical end-game of our modern day rabbit hole economics. Keynes would be weeping, and with good reason.

[Nick Rowe]

rsj: "...then propose an alternate mechanism..."

One billiard ball moves, and hits another, which moves, and hits another...Is that what you mean by "mechanism"? Economics isn't like that. People are not like billiard balls. Causality doesn't work the same way. People have expectations. If they expect everyone else (including the central bank) to act in accordance with a new equilibrium, each individual will move to that new equilibrium.

In about one month from now I will put my watch back one hour, and get up about one hour earlier by the sun. So will most Canadians. What is the "mechanism" that causes us to do that? Does the government come round and check we have all put our watches back, and fine us if we don't? Or send a cop to wake us up? Nope. I will set my watch back because I expect that everyone else will do the same. And as long as the government acts like everyone else, and doesn't falsify our expectations, the move to the new equilibrium is a self-fulfilling belief.

Given the upward-sloping IS curve, and the restoration of confidence, a successful loosening of monetary policy would, in current circumstances, probably result in a reduced demand for currency, and the central bank would accommodate that reduced demand by reducing the supply.

It's only the lack of credibility that requires a disequilibrium move by the central bank, because then it actually has to carry out its threat. Read the Swiss Bank announcement. It was all about *threatened* action. There was no action needed. Because the threat was credible. Someone looking for billiard balls will see nothing. And totally miss what happened.

[rsj]

OK, but when does economics allow you to invent channels that aren't operative?

You said that that forex purchases increase the quantity of currency which increases the price level.

But that cannot be true.

So if you still believe that CB forex purchases lead to an increased price level, you have to explain why -- what is your channel?

I proposed a channel that only works if one CB does it -- e.g. lowering forex rates stimulates exports which increases incomes (demand).

That is an orthodox channel.

You are proposing that if *all* CBs simultaneously de-value, then this will have the same effect. Obviously it must happen through some other channel. If the channel is not an increased quantity of currency -- then what is the channel?

At this point, you just said "well, it *does* work", and appealed to "confidence", without explaining confidence in what or how, or why people would have this confidence to begin with.

You still need to identify the behavioral mechanism that causes the change, instead of just asserting that the change occurs via some unknown and ineffable mechanism.

[W. Peden]

http://www.six-swiss-exchange.com/indices/security_info_en.html?id=CH0009980894CHF9

Monetary policy in action. Suddenly, corporate finance in Switzerland just got a whole lot easier.

[Nick Rowe]

rsj: "I proposed a channel that only works if one CB does it -- e.g. lowering forex rates stimulates exports which increases incomes (demand)."

First off, "channel" is still billiard ball way of thinking.

Second, you are thinking Y=C+I+G+X-M, with causality going from right to left. "Where will the demand come from?" you are asking. Because that increased demand, to move the system, seems to have to come from outside the system. It doesn't. It comes from itself.

http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/02/but-where-will-the-demand-come-from-in-praise-of-older-keynesians.html

It is the *threat* of creating an excess supply of the medium of exchange, the hot potato, that *enforces* the new equilibrium. And that ain't an unknown and ineffable force. (Not "mechanism", which is billiard ball terminology again).

[W. Peden]

Nick Rowe,

To play devil's advocate for a second, the initial effect of an exogenous increase in base money through OMOs is just to increase the cash reserves of banks. That doesn't leave the financial system unless banks withdraw it. So where's the hot potato?

[Nick Rowe]

W. Peden: What is that chart showing?

[Nick Rowe]

W Peden: were banks in equilibrium before? If so, they are not in equilibrium now.

[W. Peden]

I don't deny the explanandum, even for the sake of playing devil's advocate. I'm just looking to see if I've got the explanans right. [/pretentious philosophy of science jargon.]

As I see it, there is an important explanatory step between an increase in the monetary base and an increase in financial assets like stock markets. Banks, pension funds, insurance companies, hedge funds and other holders of securities purchased by the central bank all have a cash preference, just like any other economic agent. So they try to get back to their preferred cash-to-assets ratio by purchasing securities.

However, as a closed system, they are in a circuit. Yet, of course, there is an element here that is not simply part of the closed system: the price securities adjusts to the new supply situation and goes up as the institutions try to get towards equilibrium. So we get the effect that I observed in my first comment.

The most important and ubiquitous "leak", as I see it, is not the cash (which, as has been said, only leaves the system on the customer's demand) but on financial assets.

Tim Congdon is good on this stuff, although Sayers' book on banking from the early 1960s is still very useful. As you say, so much of awry monetary thinking is due to a fixation on interest rates, which are just a price residual from the dynamics of supply and demand (which, looked at from another angle, are the residual of price & stock!).

The reason why I think this is important is that it helps us to understand why monetary policy actions which affect the demand for base money, like IOER, are so important. Another example of monetary policy actions having important effects on demand for a kind of money would be the effects of a long term credible plan involving the reduction of broad money on the demand for broad money (which we can arguably see in the distress borrowing in Britain after the Medium-Term Financial Strategy was announced).

[K]

Nick: "It's only the lack of credibility that requires a disequilibrium move by the central bank, because then it actually has to carry out its threat."

It's only the lack of a "billiard ball mechanism" that causes the lack of credibility in the first place. QE cannot cause inflation via the quantity of money channel because there is no mechanism for reserves to become deposits. It can only work through the portfolio rebalancing channel (and then as far as I can tell, only if there is a risk of perpetual QE). Japan has done a full years GDP worth of QE and is still stuck at the ZIRB. I suspect the market has serious doubts about QE, and global competitive devaluation is clearly equivalent. I'm not certain because I still haven't decided how to evaluate the states of perpetual liquidity trap (it seems we have to consider the asymptotic equilibrium). But I'm with rsj on the need for billiard balls.

[W. Peden]

K,

"and then as far as I can tell, only if there is a risk of perpetual QE"

Permanent, not perpetual. Even this has to be softened, because bank's do not base their balance sheet composition decisions on their expectations of the monetary base 20 years' hence. If a bank acquires a bunch of non-interest bearing assets, it needs to restructure its balance sheet NOW unless it has a very good reason to anticipate a future change in monetary policy (e.g. a credible inflation targeting ceiling). Which also explains Japan.

The claim that there is no mechanism for reserves to become deposits is also inaccurate, but in an astonishingly uninteresting way.

[Nick Rowe]

K: did Japan ever announce a target? Did it back off as soon as the policy started to work? What sort of expectations does that create. A perceived temporary QE is near useless, in current circumstances. It's not what the central bank does today. It's what it threatens to do tomorrow, and the day after,...if its target is not met.

[W. Peden]

(And even then it is more technically correct to say that the base money CREATES deposits.)

[rsj]

Nick, two can play at the confidence fairy game. I say that households know that the CB is ineffectual at forcing down forex rates as the other CBs will respond. Therefore there is no confidence and you don't get a coin under your pillow.

Look, whether it is fairies, Angels, or God himself, when you need to appeal to a Deus-ex-Machina to make your case, then you have left the realm of science and have both feet planted firmly in faith.

Now you may be right, but you still need argue that you are right.

Tell me *why* households, banks, and firms, responding to a joint announcement by all the world's central banks that they will each de-value vis-a-vis the other will respond with anything other than laughter.

There may well be a reason, but inventing a false reason (e.g. an increase in currency) wont work, and appealing to fairies wont convince.

[Nick Rowe]

Just skimmed the abstract of this paper, that came in today. It says Inflation targeting central banks have fared better during the recession.

http://www.bepress.com/bejm/vol11/iss1/art22/?sending=11512

[W. Peden]

Nick Rowe,

It would be interesting to see if the paper includes banks that have basically given up on inflation targeting (like the Bank of England) in the survey. Actually, excluding the UK is likely to improve the case for inflation targeting, but not primarily because of bad monetary policy!

[K]

Nick: "did Japan ever announce a target? Did it back off as soon as the policy started to work? What sort of expectations does that create."

All of the above. So they totally failed on the communication strategy, which might, I agree, otherwise have worked if everyone believed in the mechanism of QE. But the short rate *doesn't* depend on expectations. Expectations work In setting the short rate because bank can control it *no matter what people might be expecting*. They are in fact omnipotent (and people know it). Quantity of reserves, on the other hand, can be exploded by a factor of ten and still we could enter deflation. There's no mechanical link, and Japan is evidence. So their lousy promises were backed by some random empty actions. This is why Bernanke threatens us with *secret* weapons: he knows the danger to his own credibility if he keeps threatening us with his QE pea shooter.

[edeast]

This one is interesting.
In this NK model, price flexibility is a negative during liquidity trap, when central banks lack commitment.
http://dl.dropbox.com/u/125966/zero_bound_2011.pdf

[White Rabbit]

Nick:

I think rsj has made a valid point: please tell us *why* households, banks, and firms, responding to a joint announcement by all the world's central banks that they will each de-value vis-a-vis the other will respond with anything other than laughter?

There may well be a reason, but inventing a false reason (e.g. an increase in currency) wont work, and appealing to fairies wont convince.

If today the Canadian CB announces a new overnight rate then the market will react, because it has the past experience that the CB has the right tools to achieve its interest rate target and there's an advantage in being a first mover.

If the Canadian CB announces that the Moon is made of cheese then the result will be, after a moment of shock, widespread laughter - no matter how confidently the CB announces that the Moon is made of cheese.

There's a difference between being able to threaten with action (especially if that action has already been demonstrated in the past) and between threatening with impossible to perform action.

So you need to prove that a simultaneous policy by all CBs is physically possible, before invoking the "CB threat of action causes a Pavlovian reflex in market participants" logic.

[Nick Rowe]

Suppose (falsely) everyone has expectations exactly like rsj. That there will be $0 increase in future AD, P and/or Y. I am Ben Bernanke. I commit to printing trillions and buying up every single financial asset -- government bonds, commercial bonds, stocks, and new stocks and bonds as issued, and will keep on doing this forever and ever until my NGDP target is met. At this point even rsj says, "OK, that would maybe increase future AD by $1, and increase future NGDP by $1". So now everyone starts spending more, even if it's just a small amount. But that's not an equilibrium expectation either, because we are still not at my NGDP target, so I repeat my commitment, and rsj raises his expectation another $1 when he sees that everyone else has raised their expectation too, and raised their desired spending. And so on.

Roosevelt did it. The Swiss did it. Central banks have been loosening monetary policy and creating inflation throughout history. Did history come to a full stop in 2009?

Half the US seems to be afraid of inflation. All Bernanke has to say is "Yep, I'm going to do whatever it takes to make sure your fears are justified!"

[W. Peden]

rsj,

"It will have be the same result as what happened to currency (and deposits) after the U.S. did QE"

They'll expand in volume? I see. I'm not sure what that does for your argument. Maybe you mean like when the UK did QE-

http://www.bankofengland.co.uk/statistics/fm4/2011/jul/CHART1.GIF

- oh, that doesn't work for you either. Ooops!

[W. Peden]

Don't worry, I'm sure Japan will save the case-

http://research.stlouisfed.org/publications/iet/japan/page6.pdf

- ... oh deary me! Even during an asset price recession, QE1 in Japan STILL boosted broad money growth.

[White Rabbit]

Nick wrote:

"I am Ben Bernanke. I commit to printing trillions and buying up every single financial asset"

But that is not what the Swiss have done. They have set the forex rate which is a sterilized operation.

So the point rsj has made is that your _*"The good way to fight a currency war is to devalue your own currency by making your own currency more plentiful."*_ SNB approach cannot work and thus cannot be credible, without you outlining a rational channel that makes it credible.

Consider this simple model: assume that Switzerland and Germany are the only two countries on the planet and they trade with each other. The Swiss would set the exchange rate to 1.2000 to make the Franc more plentiful and would commit to purchase Marks every time the Franc strengthens below the magic 1.2000 boundary. This forces the exchange rate to 1.2000.

What could the Germans do to devalue their currency and make it more plentiful than under the 1.2000 exchange rate? They cannot simultaneously announce a rate target different from the SNB one :-)

Some other 'global CB action' might be credible. Do you retract your current, forex based approach (the SNB approach applied to the global economy) and will you outline the one that you think would work?

[Max]

Devaluation increases demand by changing the balance of trade. In the case where devaluation is not an option, e.g. the U.S. can't devalue against the Chinese yuan, the same effect can be obtained by trade restrictions.

But the Swiss intervention I think is more "defensive" than "aggressive". They are reacting to a sudden large currency move.

[David Pearson]

"Roosevelt did it."

Roosevelt removed a binding constraint on the path of future monetary policy. As a result of that constraint, real interest rates were 12%, real wages were spiking, corporate profits negative in aggregate, and the real value of debt was also climbing. In short, everything was going in the "wrong" direction. The result of eliminating the constraint had a large impact on the confidence fairy.

Today, there is no constraint on policy. Markets are well aware that the Fed may engage in multiple future rounds of monetary easing. That is why real interest rates are not positive 12%, but negative 2%. Real wages and the real value of debt have been falling, and corporate profits are at peak and have been rising. In short, everything is moving in the "right" direction.

I can understand why Determinant says this is like the 30's: we appear to have a large output gap. Beyond that though, the situation seems very different. As Bernanke pointed out, there were pernicious effects associated with severe deflation in the 30's; these same effects are not apparent with modest inflation in 2011.

I agree with others above that Nick is proposing global coordinated QE. Like any QE program, the jury is out on whether it works or just piles up ER's. Svensson argues real depreciation of a currency is a guaranteed solution, but this cannot be done through QE, only by stand-alone competitive devaluation.

[David Pearson]

BTW, if you want to make a comparison with the 30's, I think 1937 is much more appropriate.

[Jim Rootham]

@David Pearson What do you mean by the 'right' direction, in particular with respect to real wages and corporate profits?

[Paul Friesen]

Nick, your ideas are interesting, but I don't think that even a "good" currency war will help the current situation, for the same reason that "quantitative easing" can't do much.

The basic problem for many countries and for the world as a whole is lack of demand. The situation can only be fixed if some sector somewhere can be made to demand more. In a normal recession, central banks get housing sectors to demand more by reducing the interest rate. The basic problem we are facing is that this can't work when interest rates are up against the lower bound.

I don't think that even your "good" currency war would change that. If currencies fluctuate in value relative to each other, it will move demand around from country to country, but it will be zero-sum. Sure, money will be created, but that money will just suffer the same fate as all the money created by "quantitative easing" - it will sit in the banks, which will not lend it out to companies that are facing a lack of demand. Like quantitative easing, it might not be completely useless if it reduces interest rates on higher risk and longer term loans, but it is unlikely to do much.

We need to think about how to get some part of the world economy to demand more. To me, the only reasonable solution is government stimulus. Central banks can't do more than they are already doing.

[Too Much Fed]

White Rabbit said: "Consider this simple model: assume that Switzerland and Germany are the only two countries on the planet and they trade with each other. The Swiss would set the exchange rate to 1.2000 to make the Franc more plentiful and would commit to purchase Marks every time the Franc strengthens below the magic 1.2000 boundary. This forces the exchange rate to 1.2000.

What could the Germans do to devalue their currency and make it more plentiful than under the 1.2000 exchange rate? They cannot simultaneously announce a rate target different from the SNB one :-)

Will other things (like commodities) start rising in value but not labor? Will that lead to price inflation with very little if any real GDP growth?

[rsj]

Nick,

" I am Ben Bernanke. I commit to printing trillions and buying up every single financial asset -- government bonds, commercial bonds, stocks, and new stocks and bonds as issued, and will keep on doing this forever and ever until my NGDP target is met."

Then you are replaced.

The Federal Reserve cannot purchase any financial asset. What the bank can and cannot do is defined in the Federal Reserve Act.

It can purchase obligations of the U.S. government, gold, bills of exchange, and agricultural paper. Adding that all up, you have at most about 7 Trillion or so of assets that the central bank can buy (it does not buy gold today).

After that, there is nothing left for the CB to legally buy.

Compare that to the 60 Trillion or so of U.S. assets that are out there.

Again, this is where knowledge of the institutions is important.

By the way, there is a *reason* why central banks cannot purchase "baskets of output" or "all assets", as you keep insisting.

Even the ECB, which has no risk free assets to purchase, is limited in what it can buy, both legally and effectively (e.g. politically, by the member banks that supply capital to it).

That reason is that should the CB purchase an asset that is defaulted upon, that loss of capital becomes a fiscal transfer from the ECB to the borrower.

That fiscal transfer has to be paid for by someone -- it is not paid for by the central bank, which is an intermediary between those who supply capital to it and its borrowers.

Therefore central banks can only engage in these types of transfers should the capital suppliers allow it.

In the case of the Fed, the capital supplier is the U.S. Treasury, and if the U.S. treasury wanted to make transfers (as under TARP), then it would do so itself, after permission from Congress, as spending needs to be authorized by congress and cannot occur as a result of CB policy.

So the democratic process limits the amount of fiscal policy that central banks can conduct under the guise of monetary policy.

Even if you are confused about the distinction between the two -- the legislators are not confused; which is why they have put strict limits on what central banks can and cannot buy, in order to limit any unauthorized ex-post transfers.

[Too Much Fed]

rsj, you underestimate bernanke. All he has to do is say do what I want or there will be Great Depression II, or the need for martial law, or stock prices will fall 30% to 70%. Almost all of the "people" in congress will say oh, yes we will do that right away.

When has congress ever told the fed no?

[rsj]

Peden, obviously if you measure monetary aggregates that include bank reserves, then they will increase as a result of QE.

I'm not sure what you think this proves.

You need to look at deposits held by the non-financial sector ("deposits", in the Flow of Funds classification, includes currency as a sub-component). That does not include bank reserves or bank vault cash. So go back to the drawing board if you think you can prove your point, but I've looked at this data and you will not succeed.

[jesse]

@Nick: "Did the Swiss cut interest rates to loosen monetary policy and devalue the Swiss Franc? Nope. They just announced a target, and Bingo!"

So here we have what seems an "impossible trinity" at play, and people are speculating whether markets can "overpower" the SNB. Currency futures were pricing in a non-zero (actually reasonable) chance the SNB will fail. But the trinity states that as long as the Swiss inflation rate tracks the Eurozone reasonably closely, there is no amount of money the SNB cannot sterilize. Any capital "leakage" into the Swiss economy through banking fees or whatever would cause a devaluation. At least that's how I picture it.

There have been cases in history of markets testing small fry's central banks with their various monetary policy schemes, but several economies of similar size to Switzerland's have been operational for decades with a peg, with no collapse or surrender from their central banks, so history is not in the vigilante's favour.

[rsj]

And Nick,

I would point out that in the specific case of the ECB, this -- e.g. not having risk-free assets to purchase -- is a binding constraint.

It means that *every* asset that the ECB purchases has some risk of default which needs to be covered by ECB capital, which is supplied by the member states, the largest supplier being Germany.

The ECB is a highly-leveraged bank that has already tapped the member states twice in the last year to pony up more capital.

This is why, IMO, the ECB has not expanded its balance sheet nearly as much as has the Fed.

With your model you don't even see this constraint, and can't imagine why the ECB isn't "doing more" and expanding its balance sheet even further.

Why doesn't it just load up on Greek debt? For the same reason that any other highly leveraged bank that is funded by conservative capital suppliers doesn't load up on risky debt. You keep thinking that the ECB should do more, but it cannot do more.

In the case of the Fed, as it is allowed to buy only U.S. government debt, and as the U.S. government is on the hook for the capital, there is no constraint as the capital supplier would be defaulting on itself. That's why we've set the system up this way, irrespective of how you think the CB operates.

But with the ECB, the german government is on the hook for the capital, but the ECB can buy Greek debt. There is a real risk of default, and the ECB needs to be aware of this risk in a way that the Fed does not.

This is why the ECB hasn't done more, isn't doing more, and it isn't going to do more to expand its balance sheet. That is why the ECB has been the laggard in terms of balance sheet expansion when compared to central banks that are funded by sovereign governments.

[W. Peden]

rsj,

Japanese M2 doesn't include reserves held at the central bank; it is cash in circulation plus a range of deposits. Nor does adjusted M4 in the UK or M2 & M3 in the US. The only cash reserves banks hold outside the central bank is money held in tills. There is no direct relation between an increase in the monetary base and an increase in broad monetary aggregates. It is simply incorrect to say that these aggregates are boosted in a purely technical sense by QE.

Broad money aggregates can only be boosted by an expansion of bank balance sheets, which can only occur if the liabilities of a bank increase (liabilities are equal to assets) OR if bank balance sheets are maintained as cash in circulation increases i.e. the public's demand to take money out of banks increases and it is accomodated by the financial system.

[W. Peden]

Actually, now that I think about it, "currency in circulation" excludes till money as well, doesn't it?

[W. Peden]

Hence the "QE is just an asset swap" line, which is wrong because of the word 'just'. Take that word out and it's very accurate: in itself, QE is a portfolio change of banks' (and insurance companies and pension funds &c) holdings of securities relative to their holdings of cash.

[rsj]

Peden

Here is the graph you should be looking at:

http://research.stlouisfed.org/fred2/graph/?g=26i

Btw, loan growth granger-causes growth in the base, not the other way around. And the price level granger-causes growth in the base, not the other way around.

You are not going to find a relationship in which the size of the base granger-causes more loans or an increase in the broader aggregates, as the latter will (in normal times) lead the former -- this has been known since Kydland and Prescott. Moreover, increasing the base via QE does not increase the money held by the non-financial sector.

[rsj]

Sorry, I forgot to add currency:

http://research.stlouisfed.org/fred2/graph/?g=26k

Anyways, you can see that as a result of the QE, currency continued unchanged on its merry way, as did household deposits (which include currency as a subcomponent), even though CB assets increased substantially.

The real issue here is what is the incidence of the operation. After doing the asset swap, all the sectors have different holdings of assets, and all the interest rates adjust someone so that net bond supply is zero across all sectors (including the foreign and government sector).

But *which* sector bears the brunt of the adjustment and which sector shrugs off the adjustment?

The answer to this question is obviously a behavioral and empirical one. It has to do with house the institutions operate, which institutions are allowed to take on leverage, which are more interest sensitive, etc. This is why (some of us) keep shouting at Nick that if he is going to be making monetarist arguments, he needs to understand the institutions at least enough to know what the incidence of the intervention will be.

Only if you are working in a model without a financial sector and without a foreign sector can you cheat and avoid answering this question since, by definition, the domestic non-financial sector bears all the incidence. In that case, they might indeed rush out and buy more goods if they were forced into holding more "money".

But in our world, the financial sector is the one whose asset supply is most sensitive to changes in interest rates, and so the financial sector is the one that bears the bulk of the quantity adjustments. See here.

[W. Peden]

rsj,

That graph doesn't contain a measurement of non-financial sector deposits, so I'm not sure what its relevance is supposed to be. The non-bank economy does not exist entirely of households and non-profit organsiations, so their deposits are not equal to the volume of deposits. When a bank purchases a new security (say a corporate bond) it increases the broad money supply by paying for that security with a new deposit.

"Btw, loan growth granger-causes growth in the base, not the other way around. And the price level granger-causes growth in the base, not the other way around.

You are not going to find a relationship in which the size of the base granger-causes more loans or an increase in the broader aggregates, as the latter will (in normal times) lead the former -- this has been known since Kydland and Prescott."

This is correct IF the central bank's MB operations are limited to a trade-off of its target and its LOLR function. That doesn't make increases MB necessarily endogenous; otherwise, QE would be impossible unless it corresponded with a huge boost in the demand for base money.

"Moreover, increasing the base via QE does not increase the money held by the non-financial sector."

Yes it does, though not directly. However, you're basically conceding my point: an increase in non-financial sector money (like adjusted M4) is not directly caused by an increase in the size of the monetary base. So-

"if you measure monetary aggregates that include bank reserves, then they will increase as a result of QE."

- is inapplicable here. An increase in broad monetary aggregates is caused by an increase in the volume of assets of banks either through the purchase of securities or the creation of loans.

[W. Peden]

(Of course, since exogenous increases in the monetary base are anticipated, they may begin to have an effect on financial asset prices and ergo broad money before they even begin. Expecations matter.)

[rsj]

The Fred graph shows currency not changing as a result of QE. That is for all sectors.

We know, in normal times, that the CB must increase its balance sheet in response to an increase in the demand for currency as prices rise and the economy grows, so the two go together as long as the CB does not treat currency as a target. This relationship then leads to the erroneous belief that the CB "controls" the quantity of currency.

QE is a clear example of how the CB cannot "target" currency, but must accommodate it, as there was no increase in currency.

I also included household deposit holdings exactly to point out that the incidence does not fall in this sector. It also doesn't fall on non-financial domestic businesses, but you can add those graphs to Fred if you want. I did a sectoral decomposition of bond holdings here, and its a bit of a pain, so you can repeat the operation for all sectors for currency and deposits if you wish.

Btw, the purpose of the bond decomposition was to try to estimate which sector ended up holding fewer bonds as a result of CB increasing its own holdings of bonds. And again, the household sector shrugged off the intervention with bonds, maintaining its previous level of bond holdings throughout the intervention, just as it maintained its previous level of currency and deposits.

In all cases, CB attempts to target, rather than accommodate, non-financial sector bond or currency holdings are *effectively* asset swaps with the domestic financial sector. Not 100% of the incidence falls on this sector, but the vast majority of it does.

[W. Peden]

What do you mean by "currency"? Do you mean cash minus bank reserves? Or something else?

"In all cases, CB attempts to target, rather than accommodate, non-financial sector bond or currency holdings are *effectively* asset swaps with the domestic financial sector."

Like I said. All one has to do is add some commonsense propositions about the financial sector's demand to hold cash as a proportion of assets and one understands the most important channel of OMOs.

Anyway, do you now agree that QE does not directly increase broad money aggregates?

[W. Peden]

I checked the updated graph to see what you meant by currency- I get it now, the currency component of M1.

I suspect on some of aspects of QE we agree with each other more than we agree with Nick, IF you're acknowledging than QE can increase broad money aggregates by stimulating increases in the size of balance sheets on the part of financial institutions.

[W. Peden]

In fact, let's just try and see where we agree, (since I apparentely can't get you to defend the original controversial claim that the boost in broad money was because of boosted bank cash reserves with the central bank)-

1. We agree that the monetary base is potentially exogenous i.e. the central bank can increase M0 by buying securities with new base money.

2. We agree that, in doing so, the central bank is conducting a transaction [i]within the financial sector[/i].

3. We agree that this changes the composition of banks' assets, not their volume.

Are we together so far?

[W. Peden]

(And we agree that the exceptions to 2 are not significant.)

[rsj]

Yes, QE *can* increase broad aggregates, but it is going to be hitting the financial sector -- at least given current arrangements.

But what we care about, if we believe that there is a pure quantity channel from CB balance sheet size --> price level, is the deposits of the non-financial sector. There isn't a quantity channel in that case because we have set the financial system up to be an intermediary between the central bank and the non-financial sector.

One can imagine other financial arrangements -- e.g. households only hold gold coins -- in which there would be a clear quantity channel. That is the world in which Nick Rowe operates. Other examples might be developing nations without a widely used banking system.

The powers and limitations of the central bank, and therefore the best ways of managing the economy, are contingent on how the institutions operate. In our world, it becomes more profitable for the financial sector to undo CB portfolio shifts long before households start to change their asset allocations, and so effectively most of these quantity changes never hit the non-financial sector. Interest rates clearly hit the non-financial sector. Fiscal policy clearly hits.

Exchange rates are a bit of an odd beast, because forex intervention is not just a CB operation. It requires the cooperation of Treasury as well. It's a joint effort. In the U.S., forex interventions are handled by treasury, but the recent swap agreements can be viewed as forex intervention, and were handled by the Fed.

Therefore it's not strictly true that the SNB can purchase euros in "unlimited amounts".

It can expose itself to as much forex risk as the swiss taxpayers are willing to bear, but no more. How much is a question of how that nation arranges its institutions.

[rsj]

Peden,

yes, to 3:23. Sorry, didn't see your last comment.

[W. Peden]

rsj,

"Yes, QE *can* increase broad aggregates, but it is going to be hitting the financial sector -- at least given current arrangements."

I think there will be a tendency for it to PRIMARILY "hit" the financial sector, because the distinction between cash reserves and government bonds has been reduced by current Fed policy i.e. paying interest on excess reserves. Add to that the fact that the demand for base money increases during this kind of crisis anyway and it therefore takes a lot of change in bank portfolios to make a difference to their balance sheets.

As for banks undoing central bank portfolio operations- they can't do it, strictly speaking. An individual bank can reduce its cash reserves within the financial system, but the financial system as a whole cannot (as a matter of accounting) get back to its desired cash ratios except by increasing the quantity of money. Any attempt to restore equilibrium holdings by intra-bank transactions (say swapping securities) will do nothing more than boost asset prices further, creating a "dance of the dollar" within the financial sector and increasing the pressure for the creation of new securities.

[W. Peden]

Good to see we're together so far.

Imagine a two-bank economy with the following balance sheets:

BANK 1

-- Assets --

Loans: $70

Securities: $25

Cash reserves: $3

Other assets: $2

-- Liabilities --

Deposits: $65

Other bank borrowings: $20

Shareholder's equity: $15

+

BANK 2

-- Assets --

Loans: $70

Securities: $25

Cash reserves: $3

Other assets: $2

-- Liabilities --

Deposits: $85

Shareholder's equity: $15

(So the net volume of non-financial sector deposits is $150.)

Let's also assume that the banks have their desired composition of assets and let's assume that the central bank isn't paying interest on excess reserves i.e. cash reserves are non-interest accruing assets. Now, say the central bank buys $2 worth of securities from Bank 1. Their balance sheets now look like this-

BANK 1

-- Assets --

Loans: $70

Securities: $23

Cash reserves: $5

Other assets: $2

-- Liabilities --

Deposits: $65

Other bank borrowings: $20

Shareholder's equity: $15

+

BANK 2

-- Assets --

Loans: $70

Securities: $25

Cash reserves: $3

Other assets: $2

-- Liabilities --

Deposits: $85

Shareholder's equity: $15

(The net volume of non-financial sector deposits is still $150.)

Up to this point, I think we're pretty much in agreement- all that has changed is the composition of bank 1's asset sheet.

Let's assume, for the sake of argument, that no-one is currentely willing to take out a loan at ANY interest rate. How do the banks get back to their desired proportion of cash as a percentage of assets?

[Too Much Fed]

W. Peden, by cash reserves, do you mean currency or central bank reserves? It seems to me when the central bank purchases the security (a bond?) it uses central bank reserves, not currency.

[K]

W. Peden: "How do the banks get back to their desired proportion of cash as a percentage of assets?"

Are we at the ZIRB? If not we will be immediately, since no-one will bid for excess fed funds.

Once we are at the ZIRB they don't care how much cash they've got. Doesn't cost reg cap. Earns the same as T-Bills.

Only effect of the operation is portfolio rebalancing.

[W. Peden]

Too Much Fed,

I mean both the banks' holdings of cash in their tills and their holdings of cash at the central bank.

K,

"Once we are at the ZIRB they don't care how much cash they've got. Doesn't cost reg cap. Earns the same as T-Bills."

At any stage in this economy, cash doesn't earn anything. So bank 1 still has this clump of non-interest bearing money. It wants to get back to a more profit-making portfolio. But how?

[W. Peden]

"Only effect of the operation is portfolio rebalancing."

Totally wrong: the effect is portfolio DEbalancing.

[rsj]

RIght. At zero rates, it doesn't matter, as banks are indifferent between holding the cash and zero rate bill. Therefore they do not have a cash target, they have a cash floor -- some amount of cash (really reserves) that they need for operations.

At positive rates, if there is a corridor system, then the CB is paying a positive rate on the reserves that is roughly the same as the bill rate, so again it doesn't matter from the point of view of the bank, and again we have a reserve floor, rather than a target.

Of course, it would be stupid for a bank to supply the system with excess reserves (e.g. QE) in a positive rate environment, since the CB would be draining away its capital in transfers to banks. That is not a sustainable situation, either from an accounting view or a political view.

Therefore in practice CBs ensure that there are no excess reserves whenever rates are positive, whether the system is a corridor or U.S. style system.

Only when rates are zero do banks engage in QE, and then its a type of desperate act, akin to shouting when you've lost an argument. You hope that on the margin it helps a little bit. I'm sure it may help a little bit -- it doesn't hurt, but not a whole comes of it.

This is different than the unconventional "credit easing" that the Fed did -- direct loans to credit constrained banks and brokerage houses. Even though the system as a whole may have excess reserves, if some banks are not trusted by others and therefore cannot meet their reserve requirements, then you still have a problem the CB can solve by directly lending to those specific banks.

As a by product of that credit easing, you can have excess reserves, but the economic benefit came from easing the credit constraints on individual banks, not from the aggregate CB balance sheet size. IMO, this why there was an observed real benefit to the economy and to economic aggregates from Bernanke's initial credit easing in the midst of the crisis that was not inherited by the subsequent quantitative easing efforts.

Bernanke did save us from a black hole, not by increasing the aggregate quantity of money, but by being a lender of last resort to specific credit constrained institutions.

[W. Peden]

"At zero rates, it doesn't matter, as banks are indifferent between holding the cash and zero rate bill."

What about holding the cash or holding an appreciating security?

[rsj]

"At any stage in this economy, cash doesn't earn anything. So bank 1 still has this clump of non-interest bearing money. It wants to get back to a more profit-making portfolio. But how?"

OK, everyone wants to make more money, but in practice banks can only park their reserves at the CB, or lend their reserves to another bank (who parks them at the CB).

Banks can purchase currency with reserves and then store that as vault cash, but as currency pays 0 (and has storage costs), that is not a better option than keeping the reserves on deposit with the CB.

So while I am sure that banks would love to be able to purchase a 10% yielding bill, they cannot and are forced to sit on those reserves. They do not have the option of purchasing goods and services with their reserves, nor do banks "lend reserves" to the non-financial sector.

There was a blog recently on interfluidity in which JKH and SRW discussed some strategic options that banks could use to try to reduce their individual reserve levels -- for example, chasing away depositors, lowering their lending standards, etc.

I think the conclusion that they came to was that while some slight changes could be made, in practice these are negligible. Appropriately enough, that blog was called "The Bounds of Monetary Policy on Planet Earth".

[W. Peden]

"So while I am sure that banks would love to be able to purchase a 10% yielding bill, they cannot and are forced to sit on those reserves."

Not at the level of the individual bank. Banks can purchase securities from each other (and insurance companies and pension funds &c).

However, at the level of the financial system as a whole, they cannot eliminate the cash (a dance of the dollar situation). Nor, of course, would draining deposits change the money supply.

Lowering lending standards, in the sense of loans, isn't applicable in this hypothetical economy because no-one is willing to take out a loan at any interest rate.

So I repeat the question after this little adventure into (other) imaginary economies where banks don't want to hold any amount excess cash as a proportion of their assets: how do Bank 1 and Bank 2, as a whole, get their cash-to-asset ratio back at their desired level? We both agree they basically can't get rid of it, so what can they do?

[rsj]

"What about holding the cash or holding an appreciating security?"

If a bank wants to purchase a long dated asset against a short dated liability -- say shares of IBM -- then, it can already do this even without excess reserves.

But the main costs of borrowing short and lending long are in the capital that the bank needs to set aside, and the regulatory burdens that need to be overcome.

So as nothing prevented the banks from doing this before the QE, why would QE "enable" more balance sheet expansion than was there before? An external finance premium? How much mileage can you get out of that?

QE does not supply banks with more capital. They can still borrow short at 0% and purchase a longer dated asset, regardless of how many reserves they have. It is regulation and capital requirements that prevent this from happening, not a shortage of "money". If the former is the binding constraint, then increasing reserves wont help.

How you put this all in economic model is hard, but I know how you can't do it, which is to assume that the quantity of money, rather than the costs of borrowing money, is the binding constraint.

[W. Peden]

"So as nothing prevented the banks from doing this before the QE, why would QE "enable" more balance sheet expansion than was there before?"

I don't understand the quotation marks. When did I say anything about "enabling"?

[rsj]

Sorry about the quotes. I wasn't trying to be snarky. You can replace with "lead to", "cause" ,"bring about", etc. The point being that even though banks do like to borrow short and lend long, increasing the quantity of reserves wont lead to an increase of maturity transformation. If it does not lead to increase in maturity transformation, then banks have the option of purchasing some other short, safe, liquid bill, but all of these are at zero now.

[W. Peden]

rsj,

Let's go back to the two bank economy. As we've established (on multiple occasions) the banks cannot get back to their desired cash ratio by lending it out, passing it around between each other or whatever. So, as I've set things up so far, the increase in cash reserves has debalanced the portfolio, yet the financial system as a whole cannot get rid of that cash (and I'm assuming that the public is unwilling to take out a loan under any terms).

(Recall that, regardless of the rate of interest on t-bills, banks still have a liquidity preference. As Sayers explained, banks want to get the best possible trade off between liquidity AND returns. Otherwise, I suppose they wouldn't hold securities at all but perhaps for OMOs.)

Let's assume both banks even out their cash reserves. The money supply has not yet changed. However, in the process of trying to get back to their desired ratios, the price of securities will increase. So the incentive to hold securities has increased as a result of their greater scarcity. The banks' can respond to this price signal by purchasing a new security. Security purchases create deposits. Following, say, purchases of $150 of corporate bonds, the new balance sheets are as follows-

BANK 1

-- Assets --

Loans: $70

Securities: $149

Cash reserves: $4

Other assets: $2

-- Liabilities --

Deposits: $190

Other bank borrowings: $20

Shareholder's equity: $15

+

BANK 2

-- Assets --

Loans: $70

Securities: $149

Cash reserves: $4

Other assets: $2

-- Liabilities --

Deposits: $210

Shareholder's equity: $15

(The volume of deposits is now $250. Cash reserves are back to 2% of assets.)

Now, I THINK that the price change takes care of the very good question of why the banks hadn't bought those corporate bonds already- there has been a change in the volume of securities as a result of the QE and the price rise is self-sustaining until equilibrium. However, even if it didn't, it wouldn't apply to purchases of securities from pension funds and insurance companies (who can't buy new securities with deposits and who also have a cash preference, but can only exchange the securities and bid up their price, making the raising of new securities more attractive).

Note that at no point did the volume of broad money DIRECTLY increase as a result of QE. Banks' cash reserves are not counted within the broader aggregates. When we strip away lending and bank solvency, the lingering and ubiquitous effect of QE is on non-cash financial assets. It IS about portfolios and asset swaps, but it turns out that these are very important.

(For the sake of simplicity, I've assumed a total unwillingness on the part of the public to borrow and consequentely left out the effects of increased security prices on all assets and therefore on the wealth of the public, which is obviously very important to their creditworthiness and therefore the willingness of banks to lend.)

The empirical basis for this is familiar: when we see credible QE announced, asset prices are boosted. Hopefully, the explanation of the transmission mechanism given above (so woefully lacking in most monetarist talk) turns this process from a "black box"/confidence fairy to the kind of pleasingly hydraulic process that people rightly demand when expectations &c. are invoked.

[W. Peden]

Ah! Defeated by the blasted apostrophe! The second use of 'banks' in paragraph four shouldn't be a plural possessive.

[K]

W Peden: the answer (that I assume you are looking for) is that if the bank doesn't want to sell it's security holding to the cb it has to go out and buy that security it in the market. This will create a new deposit - maybe. Most government bonds are already monetized (repo'd) since it's the cheapest form of funding available (the current on the run 10y repos around -2.5%). Certainly any bond that's held by a bank or in a margin account is fully monetized. So if the cb buys it there is no change in deposits.

But let's not confuse the conversation by mixing up money creation and risky asset purchases. Any risky asset purchase by the cb can be viewed as 1) a purchase of a t-bill plus 2) the exchange of a t-bill for risky asset (credit easing or operation twist or whatever). We can debate the merits of 2) independently if you want, and there are subtle issues for sure. But can we agree to separate the two operations? and deal with the first one first?

[W. Peden]

k,

Actually, the purchase of the $2 worth of securities by the central bank is the LEAST controversial thing here. I think we all agree that (with whatever minor provisios) the purchase of the security by the central bank doesn't (in itself) change the volume of deposits.

I think that, if we want to explain the apparent effect of QE on asset prices and broad money aggregates, we have to look at private bank portfolios and private bank money creation. If we don't have an explanation of this regularity in now several cases of QE, then we're unable to explain modern monetary policy. The fact that- at near 0% interest rates- OMOs should have ANY relation to asset prices and broad money aggregates demands an explanation.

I've scoured Sayers, Congdon, Pepper, Friedman, Goodhart and more (though Friedman and Goodhart were not really useful at all) and come to the conclusion outlined above. It's all a matter of basic accounting in the end: banks have preferences, banks have assets- how do the twain meet?

I understand the appeal of chartalism & MCT. I wouldn't go to so much effort to argue against it unless I wished it was true (I've spent even longer arguing against monetarism, by the way) and there are aspects of the monetary system that I think they get right. Ultimately, however, I must conclude that the broad money supply is exogenous, even at the ZIRB, on the basis of evidence + an a priori account of a causal process. Ditto money neutrality, where I think that the a priori case for non-neutrality is better and the empirical case against non-neutrality is much much better.

[Too Much Fed]

"How do the banks get back to their desired proportion of cash as a percentage of assets?

Posted by: W. Peden | September 09, 2011 at 04:51 PM"

Let's add a private sector, and bank 1 purchases the same security (a gov't bond) from an entity in the private sector.

If so, is the central bank hoping the private entity will stop saving and spend early? Is it also hoping that gov't interest rates will come down, other interest rates will come down, and some entity(ies) will borrow and spend (especially the lower and middle class on cars and housing)?

[W. Peden]

Too Much Fed,

Say bank 1 purchases the security from a pension fund with a new non-interest bearing deposit. Now that pension fund has an excess non-interest bearing balance and it can only resolve it by... Buying a security. And so the dance of the dollar continues, until new securities (and new deposits) are created.

BTW all, I note this was basically the rationale of the Bank of England. Congdon writes-

"But Britain’s QE experiment had a simple and persuasive rationale: that large purchases of government securities by the state (either by the central bank or the government) from the non-bank private sector create new money balances when they are financed from the commercial banks..."

[rsj]

Peden, you are going to have a hard time proving that QE boosts asset prices. First, the M3-M2 leads M2, etc., all the way on down, as no sane central bank targets quantities, they set a rate target and then accomodate whatever quantity is demanded. So then you pass to special crisis interventions, in which fiscal policy has as good if not better statistical correlation with boosting asset prices than monetary -- but neither is significant, and these markets are volatile. Land prices haven't done well in Japan, and neither has the Nikkei. Moreover there is a very simple reason why fiscal policy or interventions such as TARP would boost prices, and that is because they directly increase earnings or equity. No fairy is needed.

After failing there, you will turn to models that basically beg the question, and then try to calibrate them, etc. That is not robust evidence.

So while it may well be that government policy as a whole influences asset prices -- and it clearly does -- making the case that is not interest rates, tax policy, the current account, or deficit spending, but rather open market operations that are the most effective is going to be hard. I can't even imagine how you find a significant relationship in a model-free context, as most governments engage in massive deficit spending simultaneously with quantitative easing, and the amounts in the former case are usually much higher.

[Too Much Fed]

W. Peden said: "Say bank 1 purchases the security from a pension fund with a new non-interest bearing deposit. Now that pension fund has an excess non-interest bearing balance and it can only resolve it by... Buying a security. And so the dance of the dollar continues, until new securities (and new deposits) are created."

1) What if it buys stock with the proceeds of the gov't bond sale?

2) What if a person sells the gov't bond (at a gain) and for whatever reason decides to stop saving and buy a good/service?

[W. Peden]

rsj,

Here's an example of OMOs linking up with asset prices in a case where one can't make a case for fiscal policy being involved: look at asset prices in the US from the anticipation of QE2 after mid-2010 onwards. That was long after the fiscal stimulus was announced. Incidentally, I'm not committed to the position that OMOs are the "most effective".

Japanese asset prices aren't problematic for me either, since the BoJ has engaged in at most one credible QE programme and counterfactually you'd expect asset prices to be falling severely during that period (dot com bubble + 9/11).

As for quantities, QE in the UK was nearly 10 times the size of the fiscal stimulus. QE1 in the US was bigger than the Bush + Obama stimuluses combined, while QE2 was only a little smaller than the entire Obama stimulus. I'm NOT saying that fiscal policy is inefficacious; simply that in pure quantitative terms it was smaller than QE in the most recent crisis in the UK and US.

[W. Peden]

Too Much Fed,

1) Then the sellers of the stock have excess cash, which either temporarily passes into circulation outside the banking system or is deposited immediately in the banks, boosting cash reserves.

2) Then the seller of that good or service has excess cash &c. Incidentally, insofar as this creates new business, it increases the creditworthiness of the sellers of those goods & services.

I don't know how significant these leakages are, but the dance of the dollar is pretty mundane once the money leaves the financial system. It's only in the finacial system that the effect of the increase of the monetary base on broad money can be controversial, and even then only under special interest rate circumstances.

[David Pearson]

W. Peden,

How does QE affect asset prices? The private sector's portfolio of risk assets reflects the return they target in exchange for taking a certain amount of risk. When the Fed buys risk assets, the effect on the private sector's risk (duration, earnings volatility) is zero. That is because the Fed's balance sheet is now exposed to that risk, which means their profit remittances to the Treasury are also exposed. Treasury, in turn, passes that risk back to the private sector in the form of Treasury supply and/or higher taxes.

The Fed can only use its balance sheet to extinguish liquidity risk. All other risk it passes on to Treasury and back to the private sector.

[W. Peden]

David Pearson,

"How does QE affect asset prices?"

Initially by reducing the volume of interest bearing assets and primarily through the subsequent increase in banks' cash reserves viz. their overall portfolios, as described above.

[W. Peden]

(As I said, I think the initial price effect is sound, insofar as it appeals to the fairly uncontroversial principle of supply & demand; however, even if it wasn't, there would still be a portfolio effect via purchases from pension funds, insurance companies and other sellers of bonds to the Fed who can't finance security purchases through new deposits under any circumstances.)

[rsj]

W. Peden,

My personal opinion in the U.S. was that the stock market was saved by the combination of credit easing -- e.g. the alphabet soup lender of last resort agencies -- together with TARP.

Without these two facilities, most of the banks, which accounted for 1/4 of SP 500 market cap pre-crisis -- would have zeroed their equity holders out. To me, that is a clear example of strategic fiscal and monetary policies, working together, that significantly boosted asset prices at relatively low cost.

My second order assessment is that without our safety net spending -- unemployment insurance, welfare spending, and foodstamps, the market would have fallen much more.

In terms of amounts, the U.S. Federal government has been running deficits approximately equal to 10% of GDP in 2008, 2009, and 2009, which by far exceeds the QE amounts.

I think Britain is similar, as a recent IMF study showed that by the largest cause of deficit spending is safety net spending, not stimulus spending.

The problem is that we are dealing with small sample sizes, and once you start playing the anticipation game, then your job of explaining asset price movements becomes harder, as the whole concept of "too big to fail" suggests that the banks themselves believed that they would be bailed out and the CB would come charging to the rescue -- and yet we still had a market crash. If investors anticipated higher unemployment insurance payments, then why didn't they anticipate that the CB would engage in QE. Bernanke's speech was much discussed.

Perhaps the only *unanticipated* event in this whole crisis is that Lehman was allowed to collapse, and once investors believed in "no more Lehmans", then the market recovered.

By the way, I am not really disagreeing about whether OMO raised asset prices or not in specific cases.

But I don't think it is a reliable policy instrument.

[W. Peden]

"My second order assessment is that without our safety net spending -- unemployment insurance, welfare spending, and foodstamps, the market would have fallen much more."

Ceteris paribus, I agree 100% here. Accomodating a fiscal deficit is a very straightforward kind of monetary policy "action" for an independent central bank to take.

"My personal opinion in the U.S. was that the stock market was saved by the combination of credit easing -- e.g. the alphabet soup lender of last resort agencies -- together with TARP.

Without these two facilities, most of the banks, which accounted for 1/4 of SP 500 market cap pre-crisis -- would have zeroed their equity holders out. To me, that is a clear example of strategic fiscal and monetary policies, working together, that significantly boosted asset prices at relatively low cost."

In 2008, I agree that it was a mix of factors. In 2010, I have difficulty finding an explanation other than QE2, especially given that that was when the world headwinds (the eurozone debacle) started picking up.

"In terms of amounts, the U.S. Federal government has been running deficits approximately equal to 10% of GDP in 2008, 2009, and 2009, which by far exceeds the QE amounts.

I think Britain is similar, as a recent IMF study showed that by the largest cause of deficit spending is safety net spending, not stimulus spending."

That's the total deficit though, most of which was simply an automatic consequence of lost revenues. I don't know of any model where simple deficit spending is supposed to have a great effect; it's public fixed investment that is so important in Keynesianism.

Safety net spending is just that. It has some effect, but it follows the cycle. In general, the relationship between fiscal deficits/surpluses and AD has been astonishingly unimpressive in most countries since the early 1970s. I'd be surprised if it is a more important force in asset prices than monetary policy.

As for the crash itself, I tend to think that (in the US at least) as in the 1930s bank failures followed the crisis, they did not lead it. I also don't think QE was a sure thing at any point, nor was it widely understood (it still isn't!).

"By the way, I am not really disagreeing about whether OMO raised asset prices or not in specific cases.

But I don't think it is a reliable policy instrument."

I suspect we're moving towards some sort of agreement. All I am proposing is that, even in a highly restricted case where the deck is stacked against QE (an economy where no-one wants any new loans on any terms) it still has a technical relation to broad money and therefore aggregate demand. It may take HUGE quantities (especially if the central bank is paying interest on excess reserves) but it's not technically impossible. And there are reasons someone might prefer monetary stimulus to fiscal stimulus.

(Were it not for independent central banks, another step that really blurs the line between monetary and fiscal policy would be underfunding/debt monetization. That wouldn't be possible if Nicholas Kaldor was right and the money supply was purely demand-side-driven, but that's a theory that has the immense virtue of being empirically falsifiable and false.)

[Too Much Fed]

W. Peden said: "As for the crash itself, I tend to think that (in the US at least) as in the 1930s bank failures followed the crisis, they did not lead it."

Were there "early" bank-like failures in both cases?

[vjk]

That cannot be right:
"I think we all agree that (with whatever minor provisios) the purchase of the security by the central bank doesn't (in itself) change the volume of deposits.
"

The purchase of the security by the central bank does in fact change volume of deposits.
Operationally, it cannot be otherwise when dealing with the private sector.

Subsequent factors may be at work to dissipate the deposit increase.

[W. Peden]

Too Much Fed,

Yep, though the timing doesn't suggest a connection. The financial crisis didn't pick up big steam until October 2008.

vjk,

I think it depends on the seller. If the seller is a commercial bank, then the transaction doesn't change the volume of non-banking sector deposits.

If the seller is, say, a pension fund, then I think you're correct and this is not a minor provisio. In this case, the commercial bank operates only as an intermediary. Even Tim Congdon describes this process as "fiendishly complex", but it seems legitimate to me: one could say that, by "drawing in" the security to the financial sector and replacing it with a demand deposit, the central bank has directly changed the volume of non-banking deposits (or as close as it can come to directly changing the volume). So I'm very willing to concede a mistake here.