Is Two the new Zero? Why Divine Coincidence failed

Speculative, as Tyler Cowen would say.

When people play the Dictator Game Ultimatum Game [thanks Alex] they don't play the way economists would predict. The Nash Equilibrium is for the first player to propose a 99:1 division of the pie take-it-or-leave-it offer, and the second player to accept. Because 1% is better than nothing, so the second player will accept, and the first player knows this.

But proposals like that tend to get rejected. And 50:50 splits tend to be common. (But IIRC non-WEIRD people do play Nash, but I can't find a link.)

50% is a magic number, if there are two players. (I wonder what would happen if the Dictator Game were played with three apples, or one bottle of scotch and one bottle of beer?)

49 is not a magic number. But any number can become a magic number by custom and convention. And the 49th parallel has become a magic number for Canadians and Americans. Each would get very upset if the other crossed that line on the map. 49 has become a focal point. The 49th parallel is like a sticky price, that doesn't change at all when relative bargaining power changes a little.

0% change can also be a magic number. I won't get upset if my coffee shop raises its price by 0%.

2% per year is not a magic number. But it can become a magic number by custom and convention. I won't get upset if my coffee shop raises its price by 2% per year. Because that's the custom and convention in Canada. The Bank of Canada has been targeting 2% inflation for 20 years, and we've gotten used to it. 2% inflation has become "quasi-constitutional". If 2% is now a magic number for the Bank of Canada, it wouldn't be surprising if 2% is a magic number for all of us.

The game played between a monopolistically competitive firm and its customers is not exactly the same as the Dictator Ultimatum Game, because it's not a zero-sum game. But like the Dictator Ultimatum Game, the firm moves first, and makes a take-it-or-leave-it offer when it sets the price. The customer decides whether to accept or reject that offer. Haggling is not normal.

If 2% has become a magic number, the Short Run Phillips Curve can become rather flat at around 2% for quite a wide range. Because the representative firm's demand curve is kinked at 2% (more elastic above 2% than below), so it has two Marginal Revenue curves at the kink -- for price increases above and below 2%. "Expectations are well-anchored" as the Bank of Canada loves to say, but "expectations" can also mean "norms". And it would take a long time at a very high or very low unemployment rate before the custom and convention eventually breaks down. The vertical Long Run Phillips Curve is rather thick for a rather long time.

The Very Short Run Phillips Curve won't be exactly horizontal at 2%. There are always "special circumstances" that allow custom and convention to be set aside temporarily. And some markets are much more like auction markets where both sides can haggle over the price. Like oil prices. But auction prices can only depart so much for so long from the 2% inflation customary and conventional core.

If I am right about this, then inflation targeting sowed the seeds of its own destruction, by creating a magic number for inflation. Divine Coincidence (where minimising the deviation of inflation from target coincides with minimising the output gap) works perfectly with a Calvo Phillips Curve. Because Calvo's fairy touches firms at random, so the sample of firms that do change prices are exactly representative of the population of firms that cannot change prices and so allow output to deviate instead. A non-random fairy messes up Divine Coincidence a bit. A 2% magic number will destroy Divine Coincidence; it destroys the very signal that central banks use to tell whether monetary policy is tight or loose. Like Goodhart's Law. All it can respond to is noise.

As I said, speculative.

Comments

[marcus nunes]

"If I am right about this, then inflation targeting sowed the seeds of its own destruction"
Likely!
[Link here NR]

[Market Fiscalist]

To confirm my understanding: Is your theory that under inflation targeting firms may stop trying to set their prices to the market clearing level and will just plan to increase prices 2% since that has become the expectations and is what everyone else does. Then when there is a demand shock not many firms will reduce their prices in response so that an inflation targetting CB will be able to hit the target at below optimal employment levels.

In effect a policy intended to deal with price stickiness has ended up making prices even more sticky so that the policy stops working ?

[Old Ari]

How many years at 2% inflation does it take to reduce the value of currency by 50%?

?73/2

[Nick Rowe]

Lars: yep, inflation targeting was more or less an accident in Canada too. But it did seem to work well for a time, when keeping inflation on target also meant keeping NGDP on the (implicit) target.

Market Fiscalist: roughly, but not exactly. If your customers respond differently to small changes in prices above and below 2%, so the demand curve is kinked at a 2% increase, then it becomes profit-maximising to increase your price by exactly 2% per year, for quite a wide range of shifts in your demand curve and marginal cost curve.

[Avon Barksdale]

"When people play the Dictator Game they don't play the way economists would predict. The Nash Equilibrium is for the first player to propose a 99:1 division of the pie take-it-or-leave-it offer, and the second player to accept. Because 1% is better than nothing, so the second player will accept, and the first player knows this. But proposals like that tend to get rejected."

No Nick, you're missing the essential problem here. In a one off non-repeated game, the Nash equilibrium is as you say. But humans seldom play non-repeated games and our minds have been built around repeated human interactions. As soon as you allow for repeated games with learning, 99:1 is not the Nash equilibrium.

I find your obsession with inflation tinkering policies weird. Look, we don't really understand this system very well. I don't care what people might say, but this stuff is much harder than running a nuclear power plant. And yet many economists (at least the not careful ones) throw around these silly little stylized models and pretend that they can run the economy off them. I'm sorry, but the world is just too complicated for that kind of engineering. Sure we might get some insight, but it's not nearly enough to engineer the outcomes of millions of lives.

Imagine if we just discovered nuclear fission, but we had no idea how it really worked. Suppose I wrote some toy "rubber band" model of the decay of the Uranium atom. Would you be comfortable if I took that model and built a nuclear reactor with it next door? Probably not. Well, as someone who is an expert in particle physics and mathematical finance, I will tell you right now that engineering an economy is far, far more complicated than building a nuclear reactor.

What frightens me is the overconfident economist who thinks he knows something that he doesn't, who then manages to convince enough politicians that he has it all figured out, and ends up engineering the economy into a great depression.

Look, we don't understand this system well enough to tinker. Just set a simple, clearly understandable rule and leave it alone.

[Nick Rowe]

Avon: "No Nick, you're missing the essential problem here. In a one off non-repeated game, the Nash equilibrium is as you say. But humans seldom play non-repeated games and our minds have been built around repeated human interactions. As soon as you allow for repeated games with learning, 99:1 is not the Nash equilibrium."

Well, actually, I spent the best years of my brain on exactly that problem. Did my PhD thesis on it. It's not quite as simple as you say.

"Just set a simple, clearly understandable rule and leave it alone."

Which simple rule? What if we adopt a simple rule, and the evidence comes in that it is not working the way we thought it would? Should we have stuck with the gold standard? The money factory nuclear reactor already exists; we don't have the option of not having it. The only option is what sort of reactor to have.

[SvN]

So, Nick, suppose firms and consumers are rational, but that updating their information set is costly, so that they are rationally inattentive. Do we still get your conjectured (for frictionless rational expectations) result?

I'm wondering whether 2% inflation is enough relative to the background noise of other (relative) price adjustments for consumers and Starbucks significantly change their behaviour. When you see the price of your Christmas Spice Latte go up, do you remember when the price was last changed and apply the 2% adjustment? Or do you ask "A small price increase....is it enough to make it worthwhile for me to research the latest prices and reoptimize my behaviour?"

[Nick Rowe]

SvN: Yes, we can indeed get the same result with search costs, so that customers know the price at the firm at which they are currently shopping, but do not know competitors' prices. With everyone behaving perfectly rationally. There's a continuum of macro equilibria. I built a model exactly like that when I was young and clever. (My model is a sorta parody of Lucas 72.) But it's a very fragile result. You need to rig the assumptions exactly right to get the model to fly; make tiny changes and it can crash and burn.

" When you see the price of your Christmas Spice Latte go up, do you remember when the price was last changed and apply the 2% adjustment?"

I would never drink Christmas Spice Latte. But I confess: no, I can't exactly remember the inflation rate for a Tim Horton's small double double. And I'm not sure how sensitive my "model" (ha!) in this post is to that assumption. So maybe my "model" here suffers from the exact same fragility as my old model. Not sure.

[SvN]

"I would never drink Christmas Spice Latte."
I apologize ... I never meant to imply that you would. Mea culpa!

[Avon Barksdale]

"Well, actually, I spent the best years of my brain on exactly that problem. Did my PhD thesis on it. It's not quite as simple as you say."

I didn't say that the ultimatum games was simple, I said it was complicated by the usual expectation of repeated games. Without including the evolutionary aspect of games like these, that's when it appears that people aren't playing the game right.

The gold standard is not a rule - it has a long history, devaluations, suspensions of convertibility, default, etc. In fact, for all those who love monkeying around with monetary policy to generate stimulus, the gold standard should be right at the top of their list. You can constantly change the conversion rate, allowing "top men" at the bank to respond to every ebb and flow of the economy.

No, I am looking for a rule that removes the discretion of "top men", that forces some academic honesty and humility back into these ideas. Most of all, I don't want the government, through its central bank, constantly manipulating units to trick us into buying more or less. Just set a price level standard as a firm rule, and stop the engineering.

[Majromax]

@Nick Rowe:

I don't think you'll quite get there with a single-period model, because to my mind firms rarely fixate on 2% but instead really look at their cost structure.

Where the 2% comes in is via instruments that make the future cost structure apparent: nominal debt and commodity futures. To the extent that some firms need to set prices now based on future costs, they must build-in an assumption about inflation. Then, markets for future costs are financially mediated based on the expectation that the central bank will more or less hit its 2% inflation target.

From there, the main point of your post applies: firms expect 2% inflation so raise prices in line with it. The financial mediation story is just how firms can act with the 2% expectations built-in without thinking about it.

@Avon Barksdale:

While the economy might also be more complicated than a nuclear reactor, the economy also happens regardless of your feelings about it. If the neighbour kids insisted on building a nuclear reactor and had no idea what they were doing, your 'rubber band model of Uranium' would still be an improvement over nothing. You should know full well from engineering that "dumb" PID controllers (only very slightly more complicated than a Taylor rule) do sage work in many real control systems, even though they are wholly ignorant of the underlying processes.

[Fred Fnord]

So let me get this straight: are you saying that when you go into your favorite restaurant and see that the prices have risen, you take out your calculator and your notebook and you see if it accords with about a 2% increase per year, and if it's more you stomp out angrily?

Or do you think other people do this?

Or do you think that people just unconsciously know what the price increases of everything they buy have been over the last year? (There is a massive amount of evidence against this, but much of it could be explained away if people's conscious minds had no access to this unconscious information, I suppose.)

Or do you think this only comes into play in big-purchaser (i.e. b2b) situations, with things like raw materials (currently suffering massive deflation, by the by)?

Or...?

[Nick Rowe]

Fred: You might write like that on other blogs, but not on this one. You could have made the same point minus the stupid snark. This is a polite Canadian blog, you hoser. That's why we have a good comment section. Any repeat and you will get deleted.

Avon: we might or might not be able to get the same macro results with just standard rationality and Nash Equilibria. But if people don't play Nash in on-shot Dictator games, I'm wondering if it matters?

Majro: but it is surprising that inflation didn't fall more, for longer, during the recession.

[Market Fiscalist]

I found it easier to think about a world where the inflation target is zero , as then all price changes will mean something, and you don't have to worry about the "was that price increase above or below my expectations" bit.

In a world where firms never change price and always adjust production to match demand then inflation targeting would not work. In a world where firms only ever changed their prices for reasons unrelated to demand for their product (such as if they had discovered better production techniques, or better suppliers) then likewise inflation targeting would not work. You need at least some firms to adjust price when demand changes or you lose the signal required by IT, and you need the number of firms to be large enough so that the signal is not lost in the noise of the other random price changes.

In a world where the CB is so good at stabilizing AD that all fluctuations in demand faced by an individual firm are due to changes in relative demand for their product then the strategy of always adjusting production rather than price when faced with changed demand and only changing price due to supply-side factors would probably be the appropriate one.

It is therefore possible that if a CB gets good at stabilizing AD by stabilizing the price level then increasing numbers of firms will move to the "always adjust production" model, and eventually the signal needed by IT will break.

When the inflation target is non-zero then you have to introduce the idea of firms increasing price by (say) 2% because that is expected which is confusing - but I think story is the same if you assume that people are reasonably good at thinking in terms of real v nominal prices.

[Frank Restly]

Avon,

"What frightens me is the overconfident economist who thinks he knows something that he doesn't, who then manages to convince enough politicians that he has it all figured out, and ends up engineering the economy into a great depression."

"No, I am looking for a rule that removes the discretion of top men, that forces some academic honesty and humility back into these ideas."

So you want a rule that prevents a great depression from happening?

Rule #1 of 1 - Each family / individual must grow their own food, stitch their own clothing, construct their own shelter, and generally provide for himself / herself. Their is no trade, no barter, and no medium of exchange. There would be no deflation, no debt, and no debt deflation. There would be no unemployment because each person would work for his own livelihood.

[Avon Barksdale]

Frank Restly,

What are you talking about? Please make some sense, man!

Economics is pretty clear. In general top-down central planning does a pretty lousy job at allocating scarce resources. For all their shame a drudgery, markets do much better. Economists marvel at how markets communicate information and allow people to build better lives for themselves. Now, market failures are possible, and we do have problems with large externalities with heavy transaction costs. Sometimes we need government solutions to those problems. Standardization of weights and measures is a great idea (and so is a carbon tax). So is standardized money from the central bank. But to prevent the bank from becoming an unelected political body, it needs independence, and the price of independence is rules. Government is important and necessary, but of the limited variety and under the rule of law.

Financial markets should not care what the central banker says. In fact, the central banker should have nothing to say. When we see markets dissect every word and statement of the central banker, we are witnessing the bank morphing into a political instrument. This is not good for society. Sooner or later, the central bank will be captured by narrow political interests. Sooner or later, the central bank will make a big mistake, just like they did in the 1930s (remember the Fed caused the great depression - and no that is not a controversial statement). Let's bring back some Hayekian humility to this situation. Get the bank on a well-defined, transparent rule. However suboptimal the rule is, the fact that we've removed discretion from the equation more than makes up it.

[Frank Restly]

Avon,

You asked for a rule that would prevent a great depression. The rule that I proposed to you would do exactly that. What seems to be the problem?

[Frank Restly]

Avon,

I get what you are saying and I am just having a little fun with you. The only way an enforceable for all time monetary rule would legitimately work would be if the federal government got out of the borrowing business altogether. Otherwise the central bank (with it's rule) and the federal government (with it's fiscal objective) can be set at cross purposes or worse put into forced consolidation.

This is the point that is lost on guys like John "Mr. Rules" Taylor.

Central bank - we will always and forever lend money at 5%. Federal government - we refuse to raise taxes enough to make 5% interest payments on our outstanding debt. Can you guess who blinks first?

[Alex Field]

What you describe is not the dictator game but the ultimatum game.

[Avon Barksdale]

Frank,

No it's not lost on me. The central bank cannot absolve the government of its fiscal sins. Stick to a rule or the bank will become a political body.

[Nick Rowe]

Alex: Ooops! Fixed. Thanks. Funny how nobody else noticed that.

[ThomasH]

I don't know if inflation targeting has failed in Canada or not. It has not failed in the US because the Fed has not followed through with its target. Rather, it seems to have a 2% inflation ceiling and an independent constraint on "too large" a balance sheet.

[Avon Barksdale]

I did notice it. That's why I called it an ultimatum game in my responses to you.

[Nick Rowe]

Avon: ah, I missed that. I thought they were two names for the same game.

[Tel]

If I am right about this, then inflation targeting sowed the seeds of its own destruction, by creating a magic number for inflation.

Agreed in principle, although the Cantillon effect never goes away, trying to target zero inflation is as good a magic number as any. Nice round number is zero. Very easy to calculate too.

In order to be effective as an economic regulator, a central bank must indulge in information destruction (not necessarily by any particular method, various options are available). I believe the popular phrase is, "When it becomes serious you have to lie."

That said, the central bank can transfer wealth quite easily by merely printing money and giving that to some and not others. Requires no trickery, just requires people to put up with it. If in doubt just let me have the printing press, I'll use it wisely, promise ;-) This is of course, presuming that wealth transfer is the basic purpose of a central bank, and I'm aware that opinions might differ on that detail.

Divine Coincidence (where minimising the deviation of inflation from target coincides with minimising the output gap) works perfectly with a Calvo Phillips Curve. Because Calvo's fairy touches firms at random, so the sample of firms that do change prices are exactly representative of the population of firms that cannot change prices and so allow output to deviate instead. A non-random fairy messes up Divine Coincidence a bit. A 2% magic number will destroy Divine Coincidence; it destroys the very signal that central banks use to tell whether monetary policy is tight or loose. Like Goodhart's Law. All it can respond to is noise.

Randomness is one (well established, and well tested) method of destroying information, actually it can do a little bit more than that; if I inject a pseudo-random sequence such that I know what the sequence is, but you don't know... then it destroys information for you (looks like white noise) but not for me (looks like predictable data). Spread-spectrum radio works this way. Hopefully you can see why this might also undermine some of the concepts of mutual trust and cooperation that economists have worked so hard to encourage, especially once all parties get involved. Then again, if the injected noise is small, it's largely harmless, right? Small change either way, who cares? All part of the healthy free market process. Better yet, plausible deniability, even more useful.

The game played between a monopolistically competitive firm and its customers is not exactly the same as the Ultimatum Game, because it's not a zero-sum game. But like the Dictator Ultimatum Game, the firm moves first, and makes a take-it-or-leave-it offer when it sets the price. The customer decides whether to accept or reject that offer. Haggling is not normal.

Go to the supermarket... send students if you don't want to be caught hanging around. Prices jerk around, there's a lot of noise injection already happening out there. Haggling IS normal, it just doesn't happen quite the same way as perhaps it used to some centuries back.

[Frank Restly]

Avon,

"No it's not lost on me. The central bank cannot absolve the government of its fiscal sins. Stick to a rule or the bank will become a political body."

The "good" versus "evil" reference to government as committing "sins" is misguided (IMHO).

This is control theory not central bank = good, government = bad.

If government wants to implement some fiscal policy (taxation / spending), they should be able to do it without influence from actions taken by the central bank.
If the central bank wants to implement some form of monetary policy, they should be able to do it without influence from actions taken by the government.

Consider the impossible trinity:
https://en.wikipedia.org/wiki/Impossible_trinity

The Impossible trinity (also known as the Trilemma) is a trilemma in international economics which states that it is impossible to have all three of the following at the same time:
1. A stable foreign exchange rate
2. Free capital movement
3. An independent monetary policy

It is possible to get all three as long as fiscal policy can operate independently of monetary policy.

[Nick Rowe]

Tel: "That said, the central bank can transfer wealth quite easily by merely printing money and giving that to some and not others."

True. But since the government owns the central bank, it's the government that spends seigniorage. And it's peanuts for 2% inflation targets. Assume NGDP growth and money growth are 5%, and currency/NGDP = 5%. 5%x5% = 0.25% of NGDP. That's peanuts compared to other taxes and transfers.

[Tel]

0.25% of NGDP. That's peanuts compared to other taxes and transfers.

I looked that one up... in terms of USA/M1 we saw growth from $1.7 trillion start of 2010 up to $2.9 trillion start of 2015, let's estimate the population of that USA at 300k, we end up with $800 per person per year. OK, I accept that many taxes are higher than that, but grab someone in the street and ask if they feel like handing over $800 per year for every person in their family, they might feel disinclined.

Of course, that calculation presumes it is spread evenly, but my whole point about Cantillon effects is that this is NOT spread evenly. Suppose 1% of the population finds themselves with an extra $80000 per year in their accounts, and 99% finds themselves with nothing extra... changes the story a bit huh? I mean $80k is decent wages for someone who works for it, but pretty good money for someone NOT working for it.

Go ahead and consider USA/M2 money stock and we see $8.4 trillion start of 2010 growing up to $11.7 trillion start of 2015 which on a per-capita basis would be $2200 per year, per person. Thus, a family of 4 has "paid" $8800 per year in M2 seigniorage, and if you find a family willing to call that "peanuts" then I should offer my housekeeping and child minding services (I know, I know, don't have the "qualifications", never mind we can work something out, shhhh don't tell the union).

[Jeff]

@Tel, seigniorage comes from the central bank purchasing things using money it either printed or created with a bookkeeping entry. This is "inside money". Currency plus reserves. M1 is currency plus demand deposits at banks plus checkable deposits and banks and thrifts, plus travelers checks. No reserves. Except for currency, all of M1 and M2 are liabilities of depository institutions or of retail money funds. Reserves, which are now the biggest part of Fed liabilities, are not part of M1 or M2, except for currency.

[Nick Rowe]

Jeff is correct. Though in the olden days we used to call Central Bank liabilities "*outside* money", and God only knows where this new-fangled blogosphere swapping of the words "inside" and "outside" came from, but it is wrong and should be banned and reversed immediately. Wikipedia gets it right. So does the Palgrave.

[Tel]

Jeff, the way I see it, either the money exists, or it does not exist, but I have a lot of difficulty believing in both at the same time.

Based on the chart, seems like a few years back there was a lot less money than what there is today, and it never suddenly appeared in my hands, meaning it must have appeared somewhere else. My conclusion is that clearly there is money being created, and all money is owned by someone... thus we have certain people made wealthier in nominal terms (printing money does not create material goods) and when you do the numbers, it isn't "peanuts" either. You can call it "inside", "outside", "upside", "flipside" or anything you like, it's still money creation. I'm certainly not going to argue with you guys as to the correct name to put on these things, because I don't see that it makes a whole lot of difference.

Can this nominal wealth be used by a select few to extract real wealth from those who were not beneficiaries of the new money creation? Well, I think it can, or at least I cannot see anything to prevent that. A dollar in anyone's hand is still equivalent to any other dollar. Spending nominal wealth that you did not earn must by definition divert real wealth away from those who did earn it.

I can demonstrate this is you like, you just need to give me the printing press and I'll show everyone what it does.

If you want to go into specifically who benefits from the current setup and how, that would be long and complex... I'm not suggesting that I know exactly the answers to that anyhow. At this stage I'm still trying to get Nick past the point where he will admit that a Cantillon effect can exist at all!

[Nick Rowe]

Tel: "At this stage I'm still trying to get Nick past the point where he will admit that a Cantillon effect can exist at all!"

That's like saying you are trying to seduce a hooker.

Print $100 and spend it on:
1. Apples
2. Bananas
3. Bonds
4. Transfer payments.

Yep, 1234 are all different. But the differences between 1234 are equivalent to fiscal policy changes. I.e. 4 minus 3 is a bond-financed transfer payment.

[Tel]

Nick, when you said "I could never really get off on Cantillon effects." (Posted by: Nick Rowe | October 29, 2015 at 06:05 PM) my interpretation was that you either didn't believe in them at all, or that you had sufficiently discounted their significance as to feel comfortable just ignoring the issue.

From my perspective, Cantillon effects are probably the most important thing that central banks do, they are essentially the only leverage that exists to achieve the supposed "Keynesian stimulus" and thus fulfil the dual mandate. I would also argue that typically central banks are bad at this; but then I certainly don't support the dual mandate either. I will admit to not doing my research, reading your earlier article from 2012:

Was it: "Do Cantillon effects matter?" Or was it "Does fiscal policy matter?" I can't tell the difference. There is no difference. "Cantillon effects" are just another name for "the effects of fiscal policy".

OK. I suppose that Austrian economists believe that fiscal policy matters. I suppose it does. And monetary policy has fiscal implications, because a faster growth rate of the money supply will mean bigger seigniorage profits for the government (as long as we stay on the left side of the Laffer Curve).

Right now I did the reading, so I get that, you have a valid point that there are similarities between fiscal policy and central bank Cantillon effects. Where does that leave Milton Friedman who supported central bank policy as a substitute for fiscal policy: was he stupid, or some kind of stealthy ninja genius? Not a serious question... I'll barge ahead. Your comment above about the 4 options is much the same thing, fair enough, I didn't appreciate where you were coming from.

Next step is how is fiscal policy different to a Cantillon money-injection effect? Totally legitimate question, I'm glad you asked. First point would be transparency and strangely enough that does link back to my comment above (December 01, 2015 at 05:45 AM) about noise injection, information destruction, and more importantly selective information destruction. With fiscal policy you have a lot of mechanisms (by design!) for transparency. Suppose government decides to give a subsidy to farmers, well I can see that, I can check how much budget has been allocated, who voted for it, usually who campaigned for it, etc. Thus, when a democratic government pays off the politically connected, they do so at the cost of being seen to be handing out favours. This doesn't usually stop them, but at least potentially it might create a bit of resistance.

How to avoid resistance? Do it on the sly of course. When central banks create money they feed a little bit into the "banking system" which gets multiplied up by a huge number of private transactions (probably what Jeff was talking about above) and the final beneficiary is really quite difficult to determine. Fool enough people and you can pretend it never happened, or at least pretend that some nebulous concept called "Capitalism" deserves to be blamed. If nothing else, you might convince Bernie Sanders supporters.

Enough of my hand waving, I'll give a concrete example: let's look at student loans, which are backed by the US Government. Note that the Feds only sign off on a guarantee for the loans, well, that's not money printing is it? Actually, yes it is money printing, because although those private banks to the dirty work of creating the loans, they depend on the taxpayer backed guarantee to be able to create those loans. Clearly government has manipulated the private investment and diverted it into student loans. This is government-sponsored malinvestment right in front of your eyes, and what is the effect? Mainly the effect is to increase the price of tertiary education... and that's exactly what you would expect from a Cantillon effect... inject the money and up go the prices.

You no doubt want a second example, sure, let's look at GNMA (or "Ginnie Mae" at it's called) which offers government backed mortgage insurance, or in other words incentive for private money creation to be injected into the housing market. How did that work out? Well, we just blamed it on that old "Capitalist" economy again, and used it to leverage more regulations, and greater government power. Fooled a lot of people didn't it? Wait! Did house prices shoot up, due to the money injection, oh yeah they sure did.

Getting back to the original question "how is fiscal policy different to a Cantillon money-injection effect" second major difference is control or lack thereof. With fiscal policy you can decide who gets it. With prices, they respond in their own way, where the money goes, so prices respond, and once that tiger is loose and running through the streets, you can't decide where it goes, it just goes. Probably Milton Friedman would think this is a good thing, but personally I think we would be better off without that scatter-gun market manipulation. Again, failures get blamed on "Capitalism" when the source of the problem was government interference. Now you are facing a double problem that government can interfere in a way that has a chaotic outcome, and gets blames on arbitrary third parties... the worst of all worlds.

If government wants to interfere, by all means let them do so in a specific manner, while taking full responsibility for their actions.

[Nick Rowe]

Governments (in civilised countries) report the profits they receive from the central bank. But it's peanuts, compared to total government revenue and expenditure. (Zimbabwe aside, until it went over the top of the Laffer curve.)

[Jeff]

Nick is right. I reversed the definitions of inside and outside money. Another name for the liabilities of the central bank is "high-powered money". That term was used because, back in the good old days before 2008, banks strove to minimize their excess reserves by lending them out. If the reserve requirement was 10 percent, and the Fed injected, say 100 million in reserves by buying that much in bonds, what would happen next is that whoever got stuck with those reserves would lend them out. That in turn would create an additional 100 million in bank deposits, and now there would be only 90 million in excess reserves, because the of the reserve requirement against deposits. But there's still 90 million in excess reserves, so those get lent out, the deposits so created require another 9 million in reserves, leaving 81 million excess. And so on and so on. The original 100 million reserve injection ends up creating 1 billion in additional bank deposits, thus the term "high-powered money".

Of course, none of this applies in the post-2008 world where the Fed pays a competitive interest rate on excess reserves. In the new world, there's no reason for the banks to lend out their excess reserves, they just leave them at the Fed and earn risk-free interest. This is why Market Monetarists so dislike the payment of interest on excess reserves. Interest on required reserves is fine, as it eliminates the "reserve tax" without changing lending incentives.

[Tel]

Governments (in civilised countries) report the profits they receive from the central bank. But it's peanuts, compared to total government revenue and expenditure.

That's a bit of a trick really. Government pays interest on bonds, the Fed owns a good fraction of those bonds so it goes "Woot! We made profit, here government, have your interest payments back again." The left hand gives to the right hand, and the right hand gives it back.

It doesn't exactly add up, in the 5 years 2010-2014 inclusive, the US government paid approx $2000 billion in interest payments, and the Fed handed back profits of $420 billion... so we still have some way to go before the Fed has entirely monetized US debt, I guess China, Japan and the EU still own a few of those bonds. At any rate, the effect is that the US government appears to be paying more interest than it really does pay, because a large chunk of that interest gets recycled. Or in other words, the Fed allows government to borrow much cheaper than the nominal bond rate would indicate.

All great fun, like a juggling act that keeps the audience amused, but expansion of the M1 monetary supply is approx three times larger than those Fed profits, while expansion of the M2 money supply is two and a bit times larger than expansion in M1. That's where the real gains are coming from IMHO.