An ice cream parable for recent monetary policy events

Consider this to be thinking out loud and not a series of positive claims asserted to be necessarily true:

1. Money is non-neutral and monetary expansion helps correct for negative AD shocks.

2. At the same time, monetary expansion itself has non-neutral real effects.  Imagine that the Fed conducts all open market operations on ice cream cones.

3. As the economy recovers, there are two options.

4. First, the Fed can keep buying lots of ice cream cones forever.  Eventually the economy ends up producing too much ice cream.  Note that in the early days of the purchases, however, the economy was producing suboptimal levels of ice cream, so this was not initially much of a distortion, if it was any distortion at all.

5. The second option is that the Fed can back off its ice cream purchases now.  This will hurt AD and it also will hurt the ice cream market, although it may pre-empt a more serious blow to the ice cream market later on.  Either the Fed finds it politically impossible to buy lots of ice cream forever, or the Fed minds the eventual distortion of the ice cream market.

6. Recently the Fed suggested it might back off ice cream purchases, to see what the price of ice cream would do, so as to learn more about the market.  We observed both a negative AD effect and a severe negative price effect in the ice cream market.  Some Fed officials then tried to talk those markets back up, just as they had been talked back down.  The talk-backs failed, reflecting the wisdom of the market.

7. We now all know that slowing down the rate of ice cream purchases will be harder than we had thought.

8. This parable assumes that injection effects are important, namely where the new money goes first.  This Austrian-like view is unfashionable, has weak theoretical foundations, and violates the Modigliani-Miller theorem, but at the moment markets seem to believe it.  Should we believe it too?

8b. But wait, you Austrians shouldn’t get too gleeful.  The implied theory of the market is anti-Wicksteed, anti “supply is the inverse of the demand to hold, and vice versa,” and “flows only,” a bit like some of the early Keynesians had argued.  Uh-oh.  (Or maybe the Austrians and the early Keynesians had a bit more in common than they like to let on.)

9. You will note that the demand for ice cream also spills over into cookies, cupcakes, and kulfi.


On point (8), the key word is "important." How important? The principle that it makes *some* difference in the economy who gets the new money first seems unassailable to me--Modigliani-Miller be damned, because it is a "perfect arbitrage" argument, and we know that arbitrage can be good but it is not perfect. If the Federal Reserve expanded the monetary base by buying my house for $25 billion, I would use the money to establish a prize for the first manned flight to Mars and back. Surely the prize would draw some resources into manned flight to Mars that are not now devoted to it. But the Fed and other major central banks instead conduct open market operations using government securities, a more impersonal and predictable channel for expanding or reducing the monetary base. Doing so greatly reduces the idiosyncracy of injection effects compared to buying somebody's house for $25 billion. I'd like to see somebody try to measure the importance of injection effects through event studies or other plausible methods.

Ah, but what you get there is not a money injection, but a gift from the fed: your house is not worth 25 million in the open market. At that point it's not a money injection, but a helicopter drop. Of course the people that are handed a bunch of money in exchange for nothing are better off.

If the fed creates new money by buying your house for its market price, the fact that you got new money doesn't really affect you: It'd be the same if I bought your house with money already in circulation. The one effect the injection has is that it increases demand for houses in your neighborhood, which affects your neighbors just as much as it helps you. The bigger the pool of items the fed is purchasing at market prices, the smaller the market distortion.

How about
1) Fed announces policy and says that it will be data driven, with inflation and unemployment the major data points
2) Based on this, the market forms expectations and prices according to those expectations
3) At a time the Fed should be loosening under its policy, it discusses timing of tightening, while claiming policy is unchanged.
4) Market adjusts its expectations

And within a very short time, the ten year goes from $1.70 to $2.50.

Yeah, this was a terrible post. Just ignore the real issues of disagreement and proceed as though they've all been resolved.

This year we're seeing a massive reduction in the federal deficit. It doesn't seem like the fed took that into account much with their timetable proposal.

Does #1 mean that the Fed "fixes" the reduction in aggregate demand by effectively stealing the purchasing power that would have been spent had demand be higher via inflation and then investing it themselves?

In addition to the theft (well, "forced borrowing"), doesn't this cause a short/medium term artificial shift (relative to non-reduced aggregate demand) from purchases of consumer goods and industrial equipment to purchases of financial securities, and thus ultimately a short-term transfer from value creators producing useful goods to at best random companies, and at worst government rent-seekers and rent-seekers who sell "financial products" for the sole purpose of collecting that money?

Of course, if the investments are ultimately profitable, more money gets destroyed in the end than created, so people will get compensated by deflation, but this assumes that the inflation doesn't continue endlessly, and that the investments are indeed profitable despite the central bank having very weak incentives to select them that way.

Overall, it seems this just promotes financial rent-seeking and effectively amounts to the devastation of the value-creating economy and theft against hardworking individuals.

It looks like Prof. Cowen has confused monetary/financial effects with real economy productive effects.

The Fed's goal (as it should be) is to increase real economic output, not to cater to financial interests.

"the economy was producing suboptimal levels of ice cream": anyone who thinks it possible for anybody to know what is the optimal level of ice cream production is being very silly.

the cows disagree ... oh and spare me ZMP milk-maker spiel

Injection effects are not necessarily important. Flows may simply be signals of future interest rate policy.

The ice cream parable seems like an argument for helicopter drops, not monetary tightening. Finally, the fact that markets responded poorly to taper-talk is what you would expect when money is still tight. When money is no longer tight, markets will not respond so poorly to taper-talk because markets won't want massive inflation that will inevitably lead to tightening anyway. The markets already expect eventual tightening, so if the Fed tightens when markets expect them too, then there won't be any adverse market reaction at all.

Wouldn't this reasoning be expressed lot more succinctly using a bit of math? The problem with verbal reasoning of this sort is that results are totally different based on "how slowly", "how much", "exactly when" etc.

The post seems un-quantifiable, non-falsifiable hand-waving.

"This Austrian-like view is unfashionable, has weak theoretical foundations, and violates the Modigliani-Miller theorem - See more at:"

I don't see how it violates M-M. from Wikipedia-

"The basic theorem states that, under a certain market price process (the classical random walk), in the absence of taxes, bankruptcy costs, agency costs, and asymmetric information, and in an efficient market,"

In number 4 you state

"Note that in the early days of the purchases, however, the economy was producing suboptimal levels of ice cream"

This violates an efficient market. The market is producing below X AND there is a group that has identified this inefficiency. If these two are true there are outsized profits to be made with publicly available information and the efficient markets hypotheses is null. Austrians would also likely argue that the Fed printing actions are akin to taxation (raising prices for some while concentrating benefits to ice cream producers) and can also create asymmetrical information (certain industries have more access and knowledge of the Fed than do others).

I also doubt that the Austrian theory violates Modigliani-Miller. At constant interest rates, debt vs. equity should not matter (much). But interest rates are being manipulated down so capital structure is altered -- either way it is financed (M-M)relative to higher rates of interest (Austrian).

I'm not sure the ice cream story is a parable. A few years ago, ice cream stores were closing. A soft serve stand on the edge of town was shuttered, and two years ago Friendly's closed a number of stores (the founders of Friendly's live in our town, and the one in town was their flagship for a long time). This greatly reduced the choices for ice cream and increased the drive time to get something decent. In the last two or three years, however, we've seen two frozen yogurt stores open, two place that sell quality gelato, the re-opening of the soft-serve stand on the edge of town, and a funky ice cream parlor/cafe in the next town over. This is about the best frozen treat selection I've seen since Harvard Square in the early 1980s had Steve's, Herrell's, Emack & Bolio's, Baskin Robbin, Brigham's, and Bailey's all in less than a square mile.

We need to find out if Ben Bernanke likes ice cream.

7. We now all know that slowing down the rate of ice cream purchases will be harder than we had thought.
Given what we know about human nature, why wasn't this an initial assumption?

I think you have the parable wrong. It isn't an injection of ice cream, it is an ice cream price fixing scheme.

For various reasons the Fed wanted to keep the price of capital very low. Their fulcrum was to manipulate the yield curve on Treasuries by buying them. There is no money to be made owning Treasuries for yield, it is a pittance. They are used as collateral for some other investment strategy that has higher yield. As we learned with MF Global, the same collateral is used multiple times, leveraging it's value as collateral and also ending up multiplying it's effect. There aren't very many other assets available that are considered secure and would be accepted as collateral. So the hundreds of trillions of dollars of options, carry trade, or whatever scheme someone dreams up is based on some Treasury bonds put up as collateral. This is how the injection of capital worked.

The longer the QE policy lasted, more of the financial system became based on the assumption that QE would continue. I wonder how many VAR models assumed the continuation of the policy since it had been around for years already. We learned in 2008 that the financial system participants don't think, they calculate.

A hint of change of policy throws all the risk and VAR calculations out the window, and by consequence huge numbers of positions had to be wound up. In these situations being last in line means bankruptcy, hence we saw dramatic changes overnight. The capital in the end was used for something or other, some productive purpose, and some of it is no longer there.

The small hint of change cascaded through the market, changing the value of everything as large numbers of participants had to sell.

This is no different than some middle eastern authoritarian who to keep peace among the rabble sets the price of bread. That price is then set in stone and his head will roll if it changes. Bernanke acted like a two bit tin pot dictator, the stupid twit. He is stuck, and we will see the continuation of QE. Except now the cost of getting things under control is higher. The Treasury will start feeling the effects as well.

This won't be over until there is nothing left to lose. This is the beginnings of cracks in the central bank's absolute authority over prices of capital.

Three perspectives are missing here:
1. Ice cream insurance used to be cheap as payouts were expected to happen during good times. Now it is expensive as payouts will happen in bad times.
2. The Fed is better off now, but the sign is the opposite if you consolidate it with the ice cream treasury department.
3. For political economy reasons, principal agent problems lead to suboptimal Fed policy. As a monopolist, the Fed should not fear losses that much.

Explain: There has been a line of thought that "transmission channels" were/are broken. Based on the observation that riskless rates are Very Low (real rates negative until at least just recently.) AND credit for many tasks in the real economy Very Hard To Get.

Or perhaps transmission channels have simply changed a great deal, along lines like @derek suggests?

If the Fed is buying Lots of ice cream cones, but all of the ice cream makers are stashing the money in mattresses, and actors who previously ate ice cream with every meal have gone to eating vegetables instead, then the Fed's behavoirs will have a big effect on the ice-cream secondary market and relatively little effect elsewhere in the economy.

The analogy breaks down because of course ice-cream isn't actually fungible the way money is.

But the question implied by the post's analogy, and analysis like those above, and the "broken transmission channel" issue, suggest a new and different question:

To what extent are the gyrations of the financial markets important to the real economy over reasonable periods of time, at least away from elections?

Certainly, changes in mortgage rates are likely to have real impacts. Will these same changes affect financing costs for durable goods? That will have an impact too.

If the Fed's main lever over the cost and availabilty of borrowed-capital has changed from setting interest rates to setting the size of the pool of riskless collateral, what implications will that have for policy?

The theoretical implication should be that Fed policy should be to intervene in the market with a diversified set of assets, ideally in all products. When T-bills are not at the lower bound, T-bills probably are the best fit, otherwise progressively longer term assets. What to intervene in is different from the question of how much and when to intervene to keep ngdp growing at a stable pace that leaves room enough for growth without too much inflation, maybe something like 5%?

"at the moment markets seem to believe it"
Is there some evidence of this claim I'm unaware of?

"7. We now all know that slowing down the rate of ice cream purchases will be harder than we had thought."

Since the market currently seems to be anticipating the future purchase slowing and has placed bets accordingly, why should the slowing itself be harder than thought?

bc recent events have revealed more 'dereks' and other non standard types in the world. (perfectly fine all of them, but harder to unwind with such diverse reactions.)

It's not me, I'm not in the market. I thought that his policy was idiotic back in 2009. I'm surprised anyone is surprised at this. Japan did similar, they would act to prevent collapse of the market, then back off and be forced back in because it was going to collapse. I suspect Bernanke thought he was smarter or had deeper pockets, and would be able to manufacture a working market. What a blithering idiot.

If the Fed is printing money, the most rational thing to do is put yourself in a place where you can make money off of it. His authority is only because many have made money listening to him. Nothing more nothing less. His asset purchases are $80 billion a month, not pocket change but not much in a multi trillion dollar bond market. But he moved the market. He price fixed using the tools at his disposal, now the rabble are rioting because they can't have cheap bread.

"It’s not me, I’m not in the market." yes, you are. (financial markets are tied to the real economy and globally connected)

The mortgage backed security bond prices illustrate the price fixing characteristic of the Fed's schemes. There is no demand for MBS' at the price the Fed is willing to pay, so when they back out, or even hint at backing out, the price drops substantially. This scheme could only have come from someone well educated and a pompous fool. Let's buy MBS' and accomplish two things; one is take them off the bank balance sheets. The banks were more than eager to sell them to the Fed because they were worth substantially less than face value. Inflate housing prices by dropping interest rates on mortgages by inflating the price of mortgage bonds. Housing prices go up, consumers feel flush, spend money (by borrowing against equity???) the economy starts moving again. Worked in 2005.

These people are blithering idiots. Not only in the first idea, does anyone learn in school that price fixing doesn't work? But also in the implementation. They know, or they should, that symbolic actions from a monetary deity are only powerful if there is an unquestioned belief in the power of that deity. The longer these policies are in place, the less magic they have. Lots of folks made lots of money selling trash to the Fed. The Fed scheme has the classic characteristics of an investment scheme where at first there are big returns but as time goes on more and more pumping in of capital is required to maintain the illusion that there are returns.

What now? Was this a Bernanke has no clothes event? I don't think so, look how long it took the European market to realize that the outcomes of those world saving monthly meetings never involved someone writing a check. But to get things back into a bottle will require more than just doing what they did last week.

Well clearly Professor Cowan starts talking about Ice Cream cones and then switches to Ice Cream. Obviously this article was nothing more than Bait and Switch ;)

As a southern californian going through this heatwave, I'd recommend that there is no downside for cheap ice cream.

All of this is unnecessary. The Fed does not have to actually buy very many ice cream cones, because of the Chuck Norris effect -- as long as the Fed is credible, the knowledge that the Fed could buy any number of ice cream cones -- and will if it has to, in order to meet its goals -- will cause markets to behave with the expectation that the long-term rate of ice cream growth will be whatever the Fed says it is.

Unfortunately, what the Fed has done is simultaneously buy lots of ice cream cones while signalling long-term near-stasis in the growth of the ice cream cone supply (2%). They should either make the target more flexible so it can go higher in times of negative AD shocks, or just target a stable 4-5% increase in nominal (as measured in ice cream cones) GDP.

Comments for this post are closed