Arnold Kling  

Elves and Helicopters

10,005% Nominal GDP Growth... Unintended Consequences of Reg...

Bryan asks,

Could the central bank hit a 20,000% nominal growth target, plus or minus 1000%?

No. Certainly not next quarter. Maybe it come close if it aimed for such a target in 2014.

Consider two thought-experiments, one involving elves and one involving helicopters.

The elves change money as the unit of account. They sneak in one night and add a zero to every price, contract, and asset in the economy. You wake up and instead of being paid $30 an hour, you are paid $300 an hour. A gallon of gas costs $25 instead of $2.50. Your $20 bill is now a $200 bill. Your $100,000 stock portfolio is now a $1 million stock portfolio.

The unit of account change is purely neutral. No one's wealth has changed. No one's income has changed. No relative prices have changed. This is the same economy as yesterday, merely with another zero in all of these places.

The helicopters change money as a store of value. They drop a huge amount of money in the economy, and it lands in proportion to where it was held before. Let's say that the helicopters double the money supply, measured as cash plus checking account balances. If I was carrying $100 in cash and had $1500 in my checking account before the helicopter drop, then after the helicopter drop I have $200 in cash and $3000 in my checking account.

In classical monetary theory, the helicopters have the same sort of effect as the elves (at least in the "long run"). Relative prices will be unchanged, real economic activity will be unchanged. Only the price level will be different.

However, in my view, classical monetary theory is wrong. The change in the store of value is not neutral. The helicopter drop will disrupt people's habitual behavior. Initially, with prices unchanged, some people will feel wealthier. However, that will change as they realize that prices are rising. As they see prices rising, people will realize that money is depreciating faster than it was before, and they will try to economize on cash balances. All sorts of relative wealth effects and relative price changes will take place. We will not end up with the same real activity as before.

Let me return this discussion to Scott Sumner's thesis. As I understand it, he wants to say that the Fed should have looked around in the second half of 2008, seen that nominal GDP was looking bad, concluded that velocity was falling, and sent the helicopters out to drop money in order to offset this. I am willing to grant that a big helicopter drop would eventually have raised prices above what they otherwise would have been. But my guess is that it would have had very little effect on near-term nominal GDP, and I am not sure whether it would have any positive effect at all on near-term real GDP.

Again, I am pursuing what Bryan calls my bizarre monetary theory. If you were brought up on either saltwater or freshwater economics, you will find it quite strange.

Am I a believer in real business cycles? To me, what was called real business cycle theory in recent decades was an attempt to describe a change in the marginal product of labor. You get a productivity shock or a tax increase that reduces labor demand. Instead, I want to talk about the need to Recalculate, that is to make major changes in the allocation of resources across firms and across industries. Because the Recalculation story does not attribute macroeconomic fluctuations to changes in M or V, it is a non-monetary story, and I suppose you can call it a real business cycle if you like.

During a Recalculation, we have to observe a drop in V. Obviously, if we keep M and P constant and lower Y, then V goes down. However, I do not see the drop in V as a playing a causal role. Nor do I see an increase in M as solving the problem. In the short run, raising M will mostly result in lowering V. In the longer run, P will be higher. Along the way, what happens to Y is not clear. If the helicopters help the economy Recalculate (for example, by surreptitiously lowering real wage rates in a helpful way), then Y will be higher. If instead they simply make things more confusing and disrupt the Recalculation, then Y will be lower.

Comments and Sharing

COMMENTS (23 to date)
Todd writes:

Arnold, your view sounds both very Austrian and very sensible to me. If holding an advanced degree in economics makes it seem bizarre, I'm not sure I want the advanced degree.

Bill Woolsey writes:


Sumner's arguments do not assume money is neutral.

Your specific expamples of how money is not neutral imply that an incease in the quantity of money raise nominal expenditure.

Recalculations don't require decreases in Y or V. They will only cause this if the demand for money rises and the incease isn't accomodated. A "recalculation" can occur in the context of shrinking demands for scarce goods that are relatively less valuable and growing demands for scarce goods that are relatively more valuable. If you think about scarcity and opportunity cost, this is how it should be.

Your arguments are simply "long and variable lags." The liquidity trap is the same idea. During some "short run," the demand to hold money is rising to match the incrase in the quantity of money. This is the same thing as velocity is falling in inverse proportion to the increase in the quantity of money.

I think even Sumner would accept that this is true--maybe even for more than a few minutes. Traditional monetarists go with months and years. At least some Keynesians are very vague because they like to leave the impression it could be permanent.

David Beckworth writes:


Your monetary analysis misses an important (maybe even the key) insight from modern macro: expectations matter. If the central bank can change expections of the future price level they can cause a radical change in present spending. That was the key point of Scott Sumner. The monetary inection may, in and of itself, have a lag as you suggest, but expectations related to those injections can change spending immediately and forcefully. That is why the Fed could have stopped the decline in velocity in late 2008, early 2009. All it would need to do is introduce an explicit inflation or price level target and the market would do the rest.

Incorporating expectations in your model changes everything.

Arnold Kling writes:

I am aware of the expectations story, and I am not buying it. If the Fed had announced last fall that it was going to try to hit a nominal GDP target of X, not one thing would have changed. Nada. Zip. Zero. My barber isn't setting prices based on his expectations for next quarter's money growth.

In terms of hide-and-seek, introducing expectations doesn't mean you are getting warmer. You are getting colder.

David Beckworth writes:


Your barber may not be setting prices based on next quarter's money growth, but what about next quarter's price level growth? If your barber knows that he will be facing higher prices going forward don't you think it would affect his behavior immediately? He doesn't even have to be aware of the Fed's inflation or price level target, all he has to know see prices going up around him (because others are expecting higher future prices) and he would respond.

A good example of this is when FDR devalued the dollar and failed to sterilize gold inflows in the early 1930s. Many attribute these actions with changing deflationary expectations and fueling the initial recovery in the mid-1930s.

Matt C writes:

Bryan: Could the central bank hit a 20,000% nominal growth target, plus or minus 1000%?

Arnold: No. Certainly not next quarter. Maybe it come close if it aimed for such a target in 2014.

Which central bank are you guys talking about? The same one that kept interest rates low all through the housing bubble, steadfastly denied there was a bubble, assured us the bank system and "the fundamentals" were sound, and then, when the house of cards started to collapse, visibly panicked and demanded/pleaded/threatened for new and unprecedented powers to Save Our Economy?

They didn't know what they were doing. They still don't know what they're doing. You guys are talking about imaginary central bankers managing an imaginary model economy, not anything that relates to the real world. But please, prove me wrong:

Can you give any real world examples of a central bank trying to initiate some kind of quantified economic change, and then hitting their target?

Vasile writes:

David Beckworth: If your barber knows that he will be facing higher prices going forward don't you think it would affect his behavior immediately?

Highly unrealistic. Still, assuming for the sake of the argument that this is the case, if everyone starts demanding tomorrow prices now, there will not be enough money for it. Now. Ahora. Heute.

So expectations matter but the actual amount of cash on hands (a reality check acting now, not tomorrow) matters even more. Markets will not clear if everyone will start demanding tomorrow's (next month's, next year's) prices.

Now you tell me what good do you expect from this.

James writes:

Very well said Vasile, thank you for that explanation.

In a way the Fed's power is made possible by and at the same time limited by the fact that 90% of American's don't worry about economic forecasts and just go about their daily business the same as yesterday.

Ryan writes:

"If your barber knows that he will be facing higher prices going forward don't you think it would affect his behavior immediately?"

Which prices? The prices the barber pays for shaving cream, shampoo, and clippers? The ones he pays at the grocery store as consumer? Or the ones he pays for electricity or heating?

This is always a story of relative price changes, and this story always has a beginning. Some prices change first, then people make adjustments in their capacities as both producers and consumers, which then necessitate additional price changes and more adjustments. This appears, at least to me, to be Arnold's recalculation view. How a central bank can manipulate these relative price adjustments via monetary policy to generate their desired pattern is beyond me. All the central bank can do is start the story with new money, but once the flow from the spigot is reduced or turned off, the economy will undergo the necessary recalculation to a consumption/production pattern that is more "real"--real in the sense that it is not generate by the creation of new money "out of thin air".

B.B. writes:

The helicopter drop is odd.

Normally, we think of the drop as distributing cash randomly. Then the story you tell of rising prices and changing prices and quantitites makes sense.

But if money is "dropped" in proportion to existing cash balances, there is a positive interest rate on holding cash. If the drop was anticipated, people would want to hold more cash, causing deflation. If future such drops are anticipated, the rise in money demand would be permanent.

You leave out the expectations dimension. If the helicopter drop is a complete surprise, we have a Friedman-Schwartz type story. If it was expected well in advance (and dropped randomly), prices would move quickly, in my view, and there would be relatively little change in real activity.

Finally, unanticipated money increases have a buffer-stock. People may hold onto cash for while before spending more. That is equivalent to a drop in velocity in the short run. But velocity rebounds as people spend.

Summer may be wrong in thinking that the Fed has solid short run control over nominal GDP.

But I do agree with him that the Fed eased too slowly in 2008 during the financial crisis, probably because the oil price surge created inflation anxieties. What a mistake.

johnleemk writes:

I think the Sumner and Kling stories are reconcilable. My reading of them is that recalculation was and probably is necessary -- Sumner actually argues that it was already taking place well before the drop in NGDP.

However, Sumner also argues that the drop in NGDP, and the associated pain, was not necessary for recalculation to continue. In his view, a fall in NGDP has all sorts of knock-on effects -- the price stickiness story, etc.

The argument is not that propping up NGDP will assist recalculation or that it will create prosperity. The argument is that it will prevent the unnecessary hardship which the sectors uninvolved in recalculation will face.

You may or may not buy this story, but I think it's plausible.

Matt C:

I believe the Reserve Bank of New Zealand is often praised for its transparency, strict discipline and adherence to its targets.

8 writes:

People have an expectation of the future based on the emergence of reality-based expectation and will recalculate.

The central bank's response is to distort reality so that they will calculate according to the wishes of the central planners.

The best policy is to improve real expectations. Cut taxes and regulations.

Radford Neal writes:

A long lag before prices increase might make sense if the central bank increased the money supply surrepticiously, and people only gradually realized what was happening, as they saw prices increasing.

But we're talking about a situation where the central bank (wisely or not) wants prices to increase. So they certainly wouldn't keep the increase in the money supply secret. Indeed, they'd take out prime time advertisements telling everyone that they'd printed lots more 100 dollar bills, and that anyone who'd obtained some, and wanted to benefit from this, should buy something quick, before prices go up. Note that the central bank will already have increased the money supply at this point, so it's not a matter of making some sort of dubious committment to future expansion.

I suspect that nobody imagines such prime time advertisements because central banks, and governments generally, are terrified of the possibility that the public might actually come to understand how the fiat monetary system works. But that's only because they generally don't want prices to go up (though they may like to create more money). But if they did want prices to go up, public realization of how fiat money works could only help the process.

David Beckworth writes:

Vasile: Why assume the money supply is fixed? The banking system can increase inside money to meet the increased demand for it. Or the central bank can accommodate the increase in money demand with an increase in outside money. Either way, expectations of higher prices can lead to an endogenous increase in the money supply that in turn creates a higher price level.

Ryan: It is the price level, not relative prices that are at issue when talking about changing expectations. (That is not to say the Fed can't distort relative prices--it does a wonderful job)The Fed can most certainly increase the price level and if the market expects the Fed to do so going forward those expectations can be self-fulfilling in the present.

Bill Woolsey writes:


Why not target nominal expenditure growth?

Do you really want to say that if everyong expects a high volume of real sales, high production, high real incomes next year, that
this would not result in more nominal expenditures now?

I am sure it depends on the situation, but if the demand for money has risen because people are worried about depression, then expectations that Fed policy means that there will not be a depression in a year should have some effect on current expenditures.

Of course, if the entire problem is a drop in productive capacity, then the increase in future nominal expenditures created by current Fed policy will only impact future prices, and so the expected inflation impacting current nominal expenditures is all that is left. And, of course, if that were really the situation, what would be the point?


This entire exercise is based upon the quantity of money rising now. What do you mean, where does the money come from?

Now, if we reverse the exercise, and the Fed promised high prices or high nominal expenditure, and cut the money supply, then certainly the contradiction between their actions and words would end with the actions stopping the inflation. The demand for money can only go so low.

Similarly, if the Fed promised now or little inflation, and undertook a policy of massive money creation, I believe that the actions would win out over the words.

The argument here is simply that announcing a policy of price level (or nominal expenditure) growth and matching those words with the proper actions, can have very rapid effects.

Still, I agree with Kling in that I want to be sure there is a process by which even pig headed individuals will raise their nominal expenditures eventually.

ThomasL writes:

At the risk of sounding incredibly stupid, say we adopted Scott Sumner's plan, we'll have 3% NGDP growth in '07, '08, and '09 through monetary manipulation of one manner or another, but everyone (broker and barber alike) can bank on 3% growth this year and every year.

So? How does that fix excess consumer leverage? Excess bank leverage? How does that prevent asset bubbles from misaligned incentives like capital gains breaks, mortgage tax deductions, CRA, favorable capital requirements, &c.?

We've got 3% NGDP growth, but so what?

I must be missing part of this plan.

Matt C writes:

johnleemk, has the NZ bank ever tried to hit an NGDP target or make substantial changes to its inflation rate targets?

I don't dispute that central banks can often maintain stability once they've got it.

But the idea that the central bank can maneuver the economy around like it was . . . a helicopter, flying us around to (say) arbitrary NGDP targets as it chooses? When has this happened?

David Beckworth writes:


I actually prefer a nominal income target but thought an example using an inflation target would be easier to explain.

There is no doubt in my mind that the key objective of the Fed should be to stabilize nominal spending.

Artturi Björk writes:

The person making the influential decision is not the barber, but the consumer. If the Fed's saying that prices will be 2% higher next year and the consumers believe this they will take the price increase into account when making their consumption desicions today. This will be reflected as higher demand for goods and will result in prices rising.

Producers set prices by observing the demand, not by following the Fed.

Yancey Ward writes:

I am absolutely fascinated by the fact that no one in the comments really discussed the "Elves Hypothetical". Would this not have raised nominal GDP? Would this not have raised nominal expenditure? And wouldn't it have done all of this without affecting real output?

I am asking a rhetorical question, but why doesn't the central bank actually do this rather than pushing new money into a limited number of outlets?

Bill Woolsey writes:


Excess "consumer leverage" is fixed by consumer s paying down debt. They can do this in the context of 3% nominal growth each year.

It doesn't prevent asset bubbles (asset prices rising "too high,") or that their collapse (asset prices dropping rapidly.)

The amounts different people spend and the allocation of resources can all change with nominal income increasing each year.

Stable growth of nominal expenditure doesn't solve all problems. It just avoids the problems resulting from fluctuations in nominal expenditures. People who make bad investments still lose money. People who borrow money have to spend less to pay it back. So?

Vasile writes:

David: Why assume the money supply is fixed?

Because at the moment of the announcement it is? Most of the barber's customers are not FED. They cannot increase their money supply synchronously with the FED announcements.

Yes, as soon as the additional money (passing maybe through 20-50 hands) will reach the barber's customers, the barber will be able (and will) raise his prices. But until than, he must either keep prices at their old level or lose marginal customers.

After all what the mechanism by which an increased supply of money increases prices is that an increased supply of money reduces the marginal utility of every unit of money. (Not necessarily for all at the same time). So people value money less and do not mind to pay more for stuff.

Yancey Ward: Very good point about the "Elves scenario". It will increase NGDP, but it will hardly make Scott Sumner happy.

Bill Woolsey: The argument here is simply that announcing a policy of price level (or nominal expenditure) growth and matching those words with the proper actions, can have very rapid effects.

You need to somehow quantify, very rapid. Let's suppose that FED announces his goal of doubling the prices in 10 years. Obviously, prices will not double overnight.

Are you claiming that the prices will increase overnight (over-month) by the right amount (5%?) if FED will make a credible announcement to increase the NGDP by 5% over year? Or even by 10% if RGDP is falling? Even if peoples/businesses cash reserves are virtually unchanged?

Vasile writes:

Arnold, I like your Recalculation story, but.. are you to shy to drop a word about profits too? Even if it is a dirty word?

So, don't tell about it anyone, but I like to thinks about depressions/recalculations periods as of "Search for Lost Profitability" quests (pompous, I know), when entrepreneurs have to figure out (again) how to make profitable investments.

Well, if this is the case, as I think it is, than it became immediately obvious, that supporting unprofitable venues does not shorten the crisis, but lengthens it. First because it blurs the division between profitable and unprofitable, second because (supported) unprofitable business compete for resources with genuinely profitable ones.

Comments for this entry have been closed
Return to top